IMPLEMENTATION OF THE OSPAR REFERENCE METHOD
Introduction
OSPAR has published a document entitled ‘Oil in Produced Water Analysis – Revised Guideline on Criteria for Alternative Methods Acceptance and General Guidelines on Sample Taking and Handling – OSPAR Agreement 2006-6’. The UK guidance draws upon much of the content of the OSPAR document, as well as findings from Round Robin tests carried out in the UK. There may be some slight differences between the two documents since the OSPAR document has to be suitable for use in all OSPAR countries and this document is specific for use in the UK only. In the first instance, this document should be used but if any doubt or confusion arises then DECC should be contacted for clarification.
The new reference method requires the use of a gas chromatograph equipped with a flame ionisation detector and requires the use of n-pentane as an extraction solvent. Note that n-pentane is a highly flammable solvent with a low boiling point. Use of this solvent offshore is subject to offshore health and safety legislation.
The following points should be considered if it is intended to directly implement the OSPAR Reference Method offshore:
(a) cost and complexity of GC-FID equipment;
(b) cost of specialist laboratory personnel;
(c) level of equipment maintenance;
(d) time taken to report ‘official’ analysis results;
(e) no need to transport samples to shore for OSPAR Method; and
(f) no need to establish correlation graph for OSPAR Method.
If an operator wishes to implement the OSPAR Reference Method directly offshore, DECC would welcome such an initiative.
Some operators may prefer to use a simpler analysis method offshore that has already been correlated against the OSPAR Reference Method in an onshore laboratory. Therefore, the DECC IR analysis method (or other analysis methods) may be accepted as an ‘alternative’ analysis method but only if it is correlated against the OSPAR Reference Method.
8.2 Using the OSPAR Reference Method directly offshore
Operators who intend to use the OSPAR Reference Method directly to analyse produced water samples for dispersed oil must closely follow the procedure in the method as given in section 8 of this document.
DECC is not responsible for the technical correctness of the OSPAR Reference Method or the content of its associated guidance notes. If errors are noted or there are queries regarding the OSPAR Reference Method or its associated guidance notes, DECC will be happy to note such queries and pass them onto the relevant competent authority. In certain instances an addendum (specific to the UK) may be compiled to summarise any recommended changes to the official method. Analysts are advised to check the DECC website periodically in order to capture any updated guidance. All correspondence must take place through the mailbox.
Onshore laboratories involved in calibrating alternative methods will need to closely follow the OSPAR Reference Method detailed in section 8 & 9 of this document and they will also need to validate the method by checking parameters such as linearity, sensitivity, discrimination, etc. The validation process and calibration checks will have to be documented so that they can be inspected by DECC upon request. Onshore laboratories should note that DECC reserves the right to carry out periodic inspections of their in-house method in order to ensure that the OSPAR Reference method is being implemented correctly.
8.3 Use of alternative oil in water analysis methods to the OSPAR Reference Method
8.3.1 Acceptance criteria for alternative analysis methods
For an alternative analysis method to be accepted, DECC must be satisfied that the alternative method will produce results statistically equivalent to the OSPAR Reference Method.
8.3.2 Methodologies to demonstrate equivalent results
For analysis of dispersed oil in water concentrations using an alternative analysis method, the establishment of a valid correlation between the alternative method and the OSPAR Reference Method by using linear regression is required.
As of the 1 January 2007 only the OSPAR Reference Method is used to determine the reportable value of the dispersed oil content of produced water. Any analysis results for produced water obtained by alternative analysis methods must not be reported directly to DECC or through EEMS, but must be first correlated against the OSPAR Reference Method before being reported.
If it has been decided to use an alternative analysis method offshore, it is mandatory to establish a correlation between the alternative method and the OSPAR Reference Method. It is also mandatory to demonstrate the validity of the correlation over time in a way that can be accepted by DECC.
8.4 General approach on how to correlate an alternative analysis method
When considering the use of alternative oil in water analysis methods in conjunction with the OSPAR Reference Method, there are two variations on a method that can be used. There is not necessarily one ‘best’ method and care must be taken by analysts to select the most appropriate method for each individual situation.
Note: Methods A1 & A2 previously proposed, which involved correlation of duplicate samples analysed by both IR and the OSPAR reference method, and described in earlier versions of this document, have now been removed, as there are no operators in the UK taking this approach. The previous Methods B1 & B2 are now referred to as Method A1 & A2 respectively in this document. If operators wish to use the previous Method A1 or A2 approach they can refer to OSPAR Agreement 2006-6 guidance on the OSPAR website atand/or discuss this with DECC.
8.4.1 Method A1 - Correlation using laboratory prepared standard oil in water solutions when the alternative analysis method uses a solvent other than n-pentane.
8.4.2 Generation of Correlation Graph
Using this method, a standard oil in water stock solution is prepared in the laboratory and used to generate a series of standard solutions.
8.4.3 Correlation standards preparation
(a) Select an alternative oil in water analysis method that is likely to yield a linear response and procure the appropriate bench top analyser / instrument.
(b) Calibrate the analyser according to the manufacturer’s instructions. Note that this can be carried out either offshore or onshore.
(c) If an instrument is calibrated onshore and shipped offshore, two standard solutions (at 20% and 80% of the instrument range) used in the generation of the original calibration graph must be analysed offshore and the analysis results must lie within 95% of the calibration graph confidence intervals to confirm that the calibration graph is still valid following shipment.
(d) Procedures specified in the DECC IR method (see section 6) on how to prepare standard solutions from a stock solution should be used. Note that two stock solutions (and two series of correlation standard solutions) will have to be generated since two different solvents will be used.
(e) Fresh, crude oil samples that are representative of the blend of produced waters being treated should be selected for use.
(f) Six or more correlation standard solutions covering a concentration range from 0-80 mg/l or other specified concentration range (see 6.5.10(c)) should be prepared from the standard stock solution by using the specified back extraction technique. The concentrations of the correlation standards should be evenly spaced.
(g) The correlation standards used may be the same standards as used to calibrate the IR instrument, but must be remeasured using the new calibration graph.
(h) Preparation of standards by direct injection of oil into the solvent is not permitted. Where use of a balance is unavailable e.g. due to vessel movement, crude oil density can be used to determine the mass of oil from a volume of oil used in the stock solution. Under certain circumstances (e.g. limited laboratory facilities) this procedure may be acceptable but it would require the consent of the appropriate DECC Environmental Inspector.
(i) If it is intended to use a blank sample in the correlation graph e.g. as one of the six standards, this should be prepared as per (d) above. It is not acceptable to include a 0,0 data result (origin) unless this is the experimentally obtained result.
(j) Should the best fit line of the correlation graph result in low absorbance or IR readings producing negative oil in water concentrations, then negative correlated results should be reported as zero.
(k) The correlation standard solutions must be stored in a fridge (4°C to 8°C) and kept tightly sealed when not in use. Standard solutions should not be kept for longer than six months.
(l) A portion of each of the calibration standard solutions must accompany the calibrated alternative instrument offshore.
8.4.4 Validation, re-calibration and re-correlation requirements
Note: As a change to the previous guidance notes, the monthly validation check is to be carried out on the IR calibration graph, not the IR : GC correlation graph.
(a) Validation must be carried out once per calendar month following generation of a new calibration graph. It is not necessary to validate the calibration graph the month it is generated. Validation can be achieved by analysing one of the calibration standard solutions by the appropriate analysis method.
(b) The corresponding result is checked against the original IR calibration graph and if the result lies within the 95% confidence intervals, the calibration is considered to be valid.
(c) If the result lies outside the 95% confidence intervals, a second sample should be analysed by the appropriate analysis method.
(d) If the second result lies inside the 95% confidence intervals, the calibration is considered to be valid.
(e) If the second result lies outside the 95% confidence intervals, then a new calibration graph, and correlation graph must be prepared using fresh standards. If possible, system checks on the instrument should also be carried out to determine if the instrument is still functioning correctly.
(f) Whilst waiting for all validation test results, analysts are required to use the current calibration & correlation graphs.
(g) Where a calibration or correlation graph cannot be established, or there are circumstances that prevent a calibration or correlation graph being prepared, produced water samples must be taken and shipped to shore for analysis by the OSPAR Reference Method (GC-FID) within seven (7) days of the sample being collected, or samples refrigerated until a replacement instrument can be shipped offshore and calibrated. and the samples analysed. In this case, a DECC Environmental Inspector must be immediately notified.
(h) A new IR calibration or correlation graph must be prepared if there have been notable changes in the make up of the discharged produced water. This may be caused by, for example, new wells coming on stream. If in doubt, contact a DECC Environmental Inspector.
(i) New calibration & correlation graphs should be implemented and made available to offshore personnel as soon as possible and certainly before expiry of existing graphs. Failure to maintain valid calibration and correlation graphs must be reported to DECC as an OPPC non-compliance.
(j) A check on the equation of the correlation line for each new 6-monthly correlation graph generated should give no more than a 20% difference in the correlated GC result from the previous 6-monthly correlation graph for a given alternative analysis method result (mg/l) unless there has been a significant change in the make up of the oil in the produced water. If a change greater than 20% is found, it should be determined whether there has been a significant change in the make up of the oil in produced water and if not, a DECC Environmental Inspector must be notified.
(k) A calibration check must be carried out if the alternative instrument has been serviced, damaged, repaired or interfered with (see 8.4.3(c)). If the results do not fall within the 95% confidence interval, then a new calibration graph must be established.
(l) Calibration and system checks for the instrument used in the alternative oil in water analysis method must be carried out in accordance with either the manufacturer’s instructions or recommended analysis methods.
(m) The results of all validations must be recorded and made available to the department on request. If possible, these results should be held offshore.
8.5 Method A2 – Correlation using laboratory prepared standard oil in water solutions when the alternative analysis method uses n-pentane as the solvent
(a) Some alternative analysis methods utilise n-pentane as the solvent to extract oil from the sample solution. If the extraction procedure follows the steps of the OSPAR Reference Method (excluding the addition of any internal standard), the need for calibration of the alternative method is reduced to doing a correlation study between instruments/methods.
(b) If using Method A2, Method A1 should be followed, substituting the word n-pentane for the word solvent. Only one set of calibration standards need be prepared when using Method A2 and care should be taken to ensure that there is sufficient volume of the calibration standards available to permit calibration analysis and monthly validations for the alternative method. Care should be taken to analyse the samples as quickly as possible and to prevent loss of solvent if transporting the samples to shore.
Summary Table of Calibration / Correlation Requirements
|
Requirement |
Frequency |
# of Standards / Samples |
|
Alternative Instrument Calibration (e.g. IR) |
6 Monthly |
6 |
|
Correlation Graph |
6 Monthly |
12 (six IR & 6 GC) |
|
Calibration Graph Validation |
Monthly |
Check 1 sample |
|
Bi-Annual Samples |
6 Monthly |
See Section 11 |
8.6 Dealing with alternative method analyser instrument breakdown
(a) If the offshore alternative analyser currently in use fails, availability of a back-up or replacement analyser within 48 hours is essential as measurement of oil in produced water is a requirement for OPPC regulatory compliance.
(b) Note that alternative analysers can be calibrated either onshore or offshore using the existing calibration standards, and correlated to the OSPAR Reference Method in the same manner as the original instrument. Details of the analyser breakdown and change of instrument must be recorded in the oil in water logbook.
(c) While availability of the newly calibrated/correlated alternative analyser is awaited, if coming from onshore, produced water samples should be taken on a daily basis in line with the oil discharge permit conditions and preserved as per 5.7. These samples should be analysed as soon as the newly calibrated/correlated alternative analyser is available, which should be no longer than seven (7) days from sampling. In this instance, the appropriate DECC Environmental Inspector should be contacted.
(d) If a pollution event or suspected pollution event occurs before the replacement analyser arrives, DECC must be notified immediately by the submission of a PON1 or OPPC Non-Compliance form giving an estimate of the discharge. Samples must also be collected at the time of the event and preserved as per 5.7. The PON1 or OPPC non-compliance form can then be updated or retracted depending upon subsequent analysis of the preserved samples.
8.7 General correlation issues covering methods A1 & A2.
8.7.1 Preparation of correlation graphs and statistical analysis
Preparing correlation graphs and interpreting statistical analysis results can be difficult and often there is more than one ‘correct‘ way of doing this. It is intended that this document adhere to the best practice as outlined in the document “Preparation of Calibration Curves - A Guide to Best Practice” which has been published by the Laboratory of the Government Chemist (LGC) [3]. The LGC document sets out in a user friendly fashion the best ways of constructing calibration/correlation graphs and how best to use statistical analysis in the preparation of graphs.
8.7.2 Labelling of axis
The X-axis will normally be the OSPAR Reference Method and is assumed to represent the ‘true’ dispersed oil in water content.
8.7.3 Straight line graphs
Responses of GC instruments and instruments used in alternative analysis methods are expected to be linear within the concentration range of 0-80mg/l, or other specified concentration range. It is therefore necessary that a positive, straight-line correlation graph be constructed when correlating an alternative method against the OSPAR Reference Method.
8.7.4 Correlation graph going through the origin
The correlation line must not be forced through the origin.
8.7.5 Reporting of oil in water values to the DECC
It is important that the correct oil in water values be reported to DECC. Reporting of oil in water values should always be in terms of the OSPAR Reference Method and not in terms of an alternative analysis method.
8.7.6 Use of Accredited Laboratories
It is no longer a requirement that onshore laboratories carrying out oil in water analysis by the OSPAR reference method be UKAS (ISO17025) accredited, though it is encouraged. It is recommended that onshore laboratories work to the requirements of ISO17025 or GLP standards. Furthermore, operators who should all have an EMS in place, must assure themselves of the quality of work being carried out by contract laboratories on their behalf. Laboratories which, prior to 1 January 2010, have not carried out analysis by the OSPAR Reference method or correlation activities for offshore operators are required to contact DECC to advise them of their intention to do so.
8.7.7 Other correlation procedures
Other correlation procedures may be considered if it can be demonstrated that a valid correlation between the alternative method and the OSPAR Reference Method can be established. However, under normal circumstances DECC would expect Methods A1 & A2 listed above to be more than adequate.
9 OSPAR REFERENCE METHOD (ISO 9377-2 As Modified by OSPAR)
Note:
This method has been downloaded from the OSPAR website . It is inserted here only for convenience. Whilst DECC does not assume any responsibility for errors contained within this method, it will forward any errors or comments to the relevant OSPAR contact. Please send any feedback to the mailbox.
Please note that the page number on the content page of the original document has been omitted, as they are no longer valid in this document. Anyone who wishes to obtain a copy of the method in the original OSPAR format should visit the OSPAR website as listed above.
10 OSPAR Reference Method of Analysis for the Determination of the Dispersed Oil Content in Produced Water (Reference number: 2005-15)
Contents Page
1 Scope
2 Normative references
3 Term and definition
4 Interferences
5 Principle
6 Reagents
7 Apparatus
8 Sampling and sample preservation
9 Procedure
10 Test report
11 Precision
Annex A (informative) Example of a column
Annex B (informative) Examples of gas chromatograms of mineral oil standard and water samples
Annex C (informative) Determination of boiling range of a mineral oil from the gas chromatogram
Bibliography
Determination of the content of dispersed mineral oil in produced water using gas chromatography with flame-ionization detection
1. Scope
This document specifies the OSPAR reference method for the determination of the content of dispersed mineral oil in water by means of gas chromatography. The method is designed for produced water and other types of waste water discharged from gas, condensate and oil platforms and allows the determination of the dispersed mineral oil content in concentrations above 0.1 mg/l. The method is a modification of ISO 9377-2 in order to include the determination of certain hydrocarbons with boiling points between 98°C and 174°C.
NOTE 1: The mass concentration of animal and vegetable fat in the test sample should not exceed 150 mg/l, because at higher values the adsorption capacity of the clean‑up column packing may not be sufficient.
NOTE 2: In the case of highly polluted waste water, especially if containing a high amount of surfactants, a loss in recovery may occur.
2. Normative references
The following normative documents contain provisions, which, through reference in this text, constitute provisions of this document. For dated references, subsequent amendments to, or revisions of, any of these publications do not apply. However, parties to agreements based on this document are encouraged to investigate the possibility of applying the most recent editions of the normative documents indicated below. For undated references, the latest edition of the normative document referred to applies. Members of ISO and IEC maintain registers of currently valid International Standards.
ISO 5667‑3:2012, Water quality ‑ Sampling ‑ Part 3: Guidance on the preservation and handling of samples.
ISO 8466‑1:1990, Water quality ‑ Calibration and evaluation of analytical methods and estimation of performance characteristics ‑ Part 1: Statistical evaluation of the linear calibration function.
3. Term and definition
For the purposes of this document, the following term and definition applies.
3.1 Dispersed mineral oil content by GC-FID
The sum of the concentrations of compounds extractable with n-pentane, not adsorbed on Florisil and which may be chromatographed with retention times between those of n-heptane (C7H16) and n-tetracontane (C40H82) excluding the concentrations of the aromatic hydrocarbons toluene, ethyl benzene and the three isomers of xylene.
NOTE 3: Substances complying with this definition are long‑chain or branched aliphatic, alicyclic, and (alkyl-substituted) polycyclic aromatic hydrocarbons, with boiling points between 98 and 525 °C.
4. Interferences
Compounds of low polarity (e.g. halogenated hydrocarbons) and high concentrations of polar substances may interfere with the determination. Surface‑active substances may interfere with the extraction step.
5. Principle
The water is extracted with n-pentane. Polar substances are removed by clean-up on Florisil. The purified aliquot is analysed by capillary gas chromatography using a non-polar column and a flame ionisation detector (FID). The total peak area between n-heptane and n-tetracontane is measured. The peak areas of the aromatic hydrocarbons toluene, ethyl benzene and the three isomers of xylene are subtracted from the total peak area. The concentration of mineral oil is quantified against an external standard consisting of a mixture of two specified mineral oils, and the content of dispersed mineral oil is calculated.
It should be investigated whether and to what extent particular problems will require the specification of additional marginal conditions.
6. Reagents
All reagents shall be reagent grade and suitable for their specific purpose. The suitability of the reagents and solutions shall be checked by carrying out a blank test.
6.1 Water for the preparation of solutions, distilled water, or water from a generator of purified water capable of removing organic traces, for example using activated carbon, shall be used.
6.2 n-Pentane, C5H12
6.3 Sodium sulfate, anhydrous, Na2SO4.
6.4 Magnesium sulfate heptahydrate, MgSO4.7H2O.
6.5 Mineral acid, e.g. hydrochloric acid, c(HCl) = 12 mol/l (ρ = 1,19 g/ml).
6.6 Acetone, C3H6O
6.7 Florisil,grain size 150 μm to 250 μm (60 mesh to 100 mesh), heated to 140°C for 16 h and stored in a desiccator.
6.8 Mixture of mineral oils, Type A & Type B
Type A should show discrete peaks in the gas chromatogram. An example is diesel fuel without any additives. See EN 590 for further information. Type B should have a boiling range higher than that of type A and should have unresolved signals in the gas chromatogram. An example of this type is a lubricant without any additives, boiling range 325°C to 460°C.
6.8.1 Standard mixture
Weigh accurately equal amounts of two different types (type A and type B, both containing no additives) of mineral oil and add enough extraction solvent (6.11.2) to give a total hydrocarbon concentration of about 10 mg/ml.
6.8.2 Calibration mixture
Prepare at least five different calibration solutions by diluting aliquots of standard mixture (6.8.1) with the extraction solvent (6.11.2). The following concentrations may be suitable:
0 (blank), 0.2 mg/ml, 0.4 mg/ml, 0.6 mg/ml, 0.8 mg/ml and 1.0 mg/ml.
Higher concentrations may be advisable for other applications.
Store the calibration mixture tightly sealed in a refrigerator (4°C to 8°C). The calibration mixtures are stable for up to six months.
6.8.3 Quality Control (QC) standard
Prepare a standard solution according to 6.8.1 in acetone (see 6.6) with a mass concentration of e.g. 1 mg/ml. The exact concentration should be about a thousand times the desired application range.
NOTE 4: When using a lubricant for QC, the stock solution is easily prepared in extracting agent (6.2), which is further diluted in acetone about tenfold before spiking the QC water.
Store the solution tightly sealed in a refrigerator (4°C to 8°C). It is stable for up to six months.
6.9 Standard mixture of n‑alkanes for system performance test
Dissolve n‑alkanes with even carbon numbers (C20, C40 and at least three furthern‑alkanes) in extracting agent (6.2) to give concentrations of approximately 50 μg/ml of the individual components. It may be necessary to use a different solvent, e.g. heptane, for the first solution; in this case dilute this first solution with extracting agent (6.2).
Store the standard mixture tightly sealed in a refrigerator. It is stable for up to six months.
NOTE 5: This solution is used to verify the suitability of the gas chromatographic system for the resolution of n‑alkanes as well as for the detector response.
NOTE 6: This solution is used to give information on the retention times of the n‑alkanes to characterize the hydrocarbons in the sample.
6.10 Reference compounds
6.10.1 n-Heptane, C7H16
6.10.2 n‑Tetracontane, C40H82
6.10.3 n‑Eicosane, C20H42
6.10.4 Toluene, C7H8
6.10.5 Ethyl benzene,C8H10
o-Xylene, C8H10
6.10.7 m-Xylene, C8H10
6.10.8 p-Xylene, C8H10
6.11 Extraction solvent containing reference compounds
6.11.1 Extraction solvent stock solution
Dissolve 20 mg of n-tetracontane (6.10.2) in n-pentane (6.2). Then add 20 ml of n-heptane (6.10.1) and dilute with n-pentane to 1 000 ml.
Store the solution tightly sealed in a refrigerator. It is stable for up to six months.
NOTE 7: n‑Tetracontane is only moderately soluble in n-pentane. Slight warming or treatment with ultrasonics accelerates the dissolution process.
6.11.2 Extraction solvent standard solution
Immediately prior to use, dilute the stock solution (see 6.11.1) with tenfold extracting agent.
6.12 Stearyl stearate test solution, C36H72O2
Dissolve 200 mg of stearyl stearate in 100 ml of extraction solvent standard solution (6.11.2).
NOTE 8: This solution is used to check the efficiency of the clean‑up procedure. Store the solution tightly sealed in a refrigerator. It is stable for up to six months.
1. Apparatus
7.1 Usual laboratory glassware. Clean all glassware by the usual procedures for this type of analysis and check for purity (measurement of blank). If necessary, rinse the glassware with extracting agent (6.2) and re‑check the blank.
7.2 Gas chromatograph, equipped with a non‑discriminating injection system and a flame ionization detector. The injection system shall allow injections up to a volume of at least 50 ml.
7.3 Column for gas chromatography,fused silica, with one of the following stationary phases:
non‑polar, immobilized 100 % dimethylpolysiloxane, or 95 % dimethyl‑/5 % diphenylpolysiloxane, or modified siloxane polymer and typical dimensions:
length: 5 m to 30 m
internal diameter: 0.25 mm to 0.53 mm
film thickness: 0.25 μm to 1.2 μm
For an example, see annex B.
It is recommended to use a precolumn (e.g. 2 m, 0.53 mm i.d., deactivated fused silica).
7.4 Data system, suitable for integrating the total area of the gas chromatogram and suitable for compensating for 'column bleeding' and for re‑integrating after drawing a new baseline.
7.5 Sampling bottles, glass, with ground glass stopper, capacity 250 ml and 1 000 ml, or with screw cap, lined with PTFE (polytetrafluoroethene).
The sampling bottle shall allow direct extraction from the bottle.
7.6 Centrifuge.
7.7 Centrifuge tubes, of capacity 100 ml, with suitable (screw)cap.
7.8 Microseparator, or other suitable device for phase separation.
7.9 Clean‑up columns, made from glass, with frit of sinter porosity 2, for example see Figure A.1.
7.10 Magnetic stirrer with bar, of length suitable to ensure thorough mixing.
2. Sampling and sample preservation
Sampling and storage shall be in accordance with ISO 5667‑3 and other OSPAR documents[1].
Fill the sampling bottle (7.5) to approximately 90 % with the sample, seal tightly and weigh (m1). Keep the sample at about 4°C and extract the sample as soon as possible, but in any case within four days.
If necessary, preserve the sample by acidifying with mineral acid (6.5) to pH 2.
NOTE 9: If formation of emulsions or a concentration of animal and vegetable oils > 150 mg/l is expected; it is advisable to withdraw additionally a smaller sample volume in a 250 ml sampling bottle.
3. Procedure
9.1 Blank test
Carry out blank tests with each series of tests in accordance with 9.3, including all reagents and glassware in the same way as the samples.
9.2 Determination of the recovery
Determine the recovery at regular intervals, preferably in each series of tests, using 900 ml of water (6.1) to which 1,0 ml of the QC standard (6.8.3) has been added. Perform the test starting with 9.3 and calculate the recovery. Ensure that the recovery is between 80 % and 110 %
9.3 Extraction procedure
Cool the sample to about 10°C, if necessary, to prevent losses of the extracting agent by volatilization.
Acidify the sample to pH 2 by adding mineral acid (6.5), if this has not been done in the field (see clause 8).
Add about 80 g of magnesium sulfate (6.4) per 900 ml of sample to avoid emulsions.
NOTE 10: The addition of magnesium sulfate is not necessary if samples are known not to form emulsions.
Add 50 ml of extraction solvent standard solution (6.11.2) and a magnetic stirrer bar, close the bottle and stir vigorously for 30 min on the magnetic stirrer (7.10).
Remove the stopper and replace it by the micro separator (7.8).
Add enough water (6.1) to allow withdrawal of the extracting‑agent layer from the micro separator, transfer it to the clean‑up column (7.9) and proceed according to 9.4.
When transferring the organic phase to the clean‑up column, take care to avoid the transfer of water, as this will incrust the surface of the sodium sulfate. It is recommended to transfer the organic layer in several steps using a pipette, or, when using the micro separator (7.8) (Figure A.2), to position the meniscus below the cock.
In the case of strong emulsions, centrifuge the extract as follows. Transfer the extracting‑agent phase together with the emulsion into a 100 ml centrifuge tube (7.7) and close the tube. Break the emulsion by centrifuging the extract.
NOTE 11: Centrifugation for 10 min to 15 min is usually sufficient.
Extraction by shaking may be adequate, but verify the efficiency.
9.4 Clean‑up procedure
Transfer the n-pentane phase (see 9.3) to a small column (see Figure A.1) filled with 2 g of Florisil (6.7) and covered with a layer of 2 g of sodium sulfate (6.3).
NOTE Pre-rinsing of the column with a few milliliters of n-pentane (6.2) can be useful to prevent the formation of channels.
Let the n-pentane phase (see 9.3), followed by an additional 10 ml of n-pentane (6.2), run through the column into a 50 ml volumetric flask until the liquid reaches the 50 ml mark.
NOTE 12: Other procedures using the same amount of Florisil for cleaning the extract, e.g. shaking the total extract with 2 g of Florisil during 30 min on a shaking apparatus, may be used as an alternative, provided the results are equivalent to the column testing of Florisil (see 9.6).
9.5 Concentration
External apparatus for concentrating the purified n-pentane extract is not allowed. The gas chromatograph (7.2) is equipped with an injection system that allows an injection of at least 50 ml of n-pentane. Injection of 50 ml is only required for low mineral oil contents, i.e. at levels near the limit of determination. Depending on the range of concentrations normally measured, a fixed injection volume, e.g. 25 or 50 ml, shall be used both for the calibration and for the measurements of the extracts.
Leave the empty sample bottle to drain for 5 min. Close the bottle with the cap used previously and determine its mass (m2) to an accuracy of 1 g.
9.6 Suitability testing of Florisil
Check the suitability of Florisil at regular intervals and each time a new batch of dried Florisil is used, as follows:
For this purpose, use a test solution of stearyl stearate (6.12) and a calibration solution of mineral oils (6.8).
Perform the clean‑up procedure (9.4) with 10 ml of the stearyl stearate solution, then add extracting agent (6.11.2) to a volume of 25 ml. Transfer an aliquot of the purified solution to a septum vial and analyze by gas chromatography (9.7). Measure the peak areas of stearyl stearate after Florisil treatment. Dilute 0,5 ml of stearyl stearate solution with extracting agent (6.11.2) to 25 ml and analyze by gas chromatography. Calculate the ratio of the peak areas for stearyl stearate in the treated and in the untreated solutions. This ratio should be less than 1. If not, activate the Florisil according to 6.7.
Perform the clean‑up procedure (9.4) with 10 ml of 2 mg/ml calibration solution of mineral oils (6.8), then add extraction solvent standard solution (6.11.2) to a volume of 25 ml. Transfer an aliquot of the purified solution into a septum vial and analyze by gas chromatography.
Determine this mineral oil recovery on the basis of the peak area between C10 to C40 in the treated (with Florisil) and untreated calibration solutions. It should be at least 80 %. If this criterion is not met, wash out the Florisil batch with ample water and activate the Florisil as given in 6.7. If the repeated test shows again that the criterion is not met, try another batch of Florisil with a different batch number.
9.7 Determination by gas chromatography
9.7.1 Adjusting the gas chromatograph
Select a capillary column with one of the stationary phases specified (see 7.3 and annex B) for gas chromatographic analysis. Adjust the gas chromatograph until optimum separation is obtained. The peaks in the gas chromatogram of the standard mixture of n‑alkanes (6.9) shall be baseline‑separated. The relative response (peak area) of n‑tetracontane (C40H82) compared with n‑eicosane (C20H42) should be at least 0,8. If not, the discrimination of the injection system is too high and the injection system shall be optimized or replaced.
9.7.2 Calibration
9.7.2.1 General
For calibration a distinction is made between initial calibration, working calibration and checking of the validity of the calibration curve. Initial calibration determines the working range and the linearity of the calibration function according to ISO 8466‑1. Perform this calibration when the apparatus is used for the first time.
In the next step establish the final working range and perform the routine calibration. Repeat this calibration after maintenance (e.g. replacement of the capillary column), after repair of the gas chromatographic systems, and in case the system has not been in use for a longer period of time, or if the validity criteria cannot be met. Check the validity of the initial calibration with each series of samples to be analyzed.
Correct all chromatograms for column bleeding, for this purpose run blank chromatograms (chromatograms with solvent only) and use these for baseline correction.
9.7.2.2 Initial calibration
Establish the preliminary working range by analyzing at least five dilutions of the calibration standard mixture (6.8). Test for linearity in accordance with ISO 8466‑1.
9.7.2.3 Routine calibration
After examining the final working range, analyze a minimum of five dilutions of the standard calibration mixture (6.8). Calculate a calibration function by linear regression analysis of the corrected peak areas. The actual sensitivity of the method may be estimated from the calculated regression function.
9.7.2.4 Check of the validity of the calibration function
Check the validity of the calibration function from the routine calibration with each batch of samples by analysis of one standard solution after every ten samples. The concentration of this standard solution shall lie between 40 % and 80 % of the working range. Make sure that the individual results do not deviate by more than 10 % of the working calibration line. If this criterion is met, assume the calibration to be valid. If not, recalibrate in accordance with 9.7.2.3.
For large batches of samples the number of analyses of the standard solution may be reduced, provided that at least three measurements are obtained for calculating the mean result.
9.7.3 Measurement
Measure the sample, the calibration solutions and the blank solution in the gas chromatograph.
At regular intervals record a gas chromatogram (blank chromatogram) by injecting extracting agent (6.2) and analyze under the same conditions as for the sample. Use a chromatogram of extracting agent to correct the areas of the chromatograms of the samples.
NOTE 13: An increase in "column bleeding" may indicate contamination of the injection system or the column.
9.7.4 Integration parameters
Integrate the gas chromatogram between n-heptane and n-tetracontane. Start the integration just after the n-heptane peak at the signal level in front of the solvent peak (S in Figure B.3). End the integration just before the beginning of the n-tetracontane peak on the same signal level (E in Figure B.3).
Check all chromatograms visually to ensure correct integration. Draw a straight line from S to E. Mark the beginning and end of the integration on the chromatogram.
Identify the peaks of the aromatic hydrocarbons toluene, ethyl benzene and the three isomers of xylene. Integrate these peaks separately and subtract their peak areas from the total integrated area.
NOTE 14: For examples of chromatograms, see annex B.
The presence of peaks between solvent peak and n‑heptane indicates that the sample probably contains low-boiling, volatile hydrocarbons.
Discrete peaks or an increased level of the baseline at the end of the chromatogram (retention time greater than that of n‑tetracontane) indicate that the sample probably contains hydrocarbons with a high boiling point. This should be mentioned in the test report.
NOTE 15: The range of the carbon numbers of n‑alkanes present in the sample is determined by comparing the gas chromatogram of the sample extract with that of n‑alkane standard solution (6.9). The corresponding boiling range can be derived from annex C.
9.8 Calculation
Calculate the hydrocarbon oil index using the equation:
(Am‑b) . f . V . w
ρ = -------------------
a . (m1 ‑ m2)
where
ρ is the content of dispersed mineral oil, in milligrams per liter;
a is the slope of calibration function, in liters per milligram;
Am is the integrated peak area of the sample extract, in instrument‑dependent units;
f is any dilution factor of the sample extract;
m1 is the mass of the filled sampling bottle, in grams;
m2 is the mass of the empty sample bottle, in grams;
w is the density of the water sample, in grams per millilitre (for fresh water 1,00 g/ml may be used);
V is the volume of the final extract, in milliliters;
b is the intercept of the y‑axis in instrument‑dependent units.
9.9 Expression of results
Express the concentration of mineral oil in water as hydrocarbon oil index, in milligrams per litre, to two significant figures.
Examples
Dispersed mineral oil 15 mg/l
Dispersed mineral oil 2.9 mg/l
4. Test report
The test report shall refer to this document and contain the following details:
- a) identity of the sample;
- b) the content of dispersed mineral oil, in milligrams per liter;
- c) any peculiarities observed during the test;
- d) any actions not specified in this document, which may have influenced the result. In addition, the following qualitative information can be provided in the test report:
- e) the boiling range of the mineral oil detected on the basis of the relative retention time versus the boiling point of the calibration mixture ofn‑alkanes;
- f) any presence of volatile hydrocarbons;
- g) any presence of hydrocarbons with a high boiling point.
5. Precision
At the moment precision and accuracy data from an interlaboratory trial are not available.
The instrumental part of this method has been checked in a small ring test.
Four laboratories in four countries (Denmark, the Netherlands, Norway and the United Kingdom) have analysed two standard solutions of hydrocarbons in the range C7 – C40 with different levels of toluene, ethyl benzene and xylenes. Each laboratory analysed the standards in triplicate on two different days.
The results are given in Table 11.1, expressed as the ratio between the measured and the true values
Table 11.1: Ring test results for the OSPAR reference method
|
|
Day |
Lab 1 |
Lab 2 |
Lab 3 |
Lab 4 |
|
Low TEX |
1 |
0.95 |
1.01 |
0.98 |
1.21 |
|
Low TEX |
2 |
0.93 |
1.01 |
0.87 |
1.09 |
|
High TEX |
1 |
0.96 |
1.06 |
1.01 |
1.35 |
|
High TEX |
2 |
0.92 |
1.02 |
0.90 |
1.13 |
In addition the laboratories were asked to determine also the content of mineral oil according to ISO 9377-2, i.e. only the hydrocarbons in the range C10 – C40. These results are given in table 11.2, expressed as the ratio between the measured and the true values.
Table 11.2: Ring test results for ISO 9377-2
|
|
Day |
Lab 1 |
Lab 2 |
Lab 3 |
Lab 4 |
|
Low TEX |
1 |
0.97 |
1.04 |
1.06 |
1.31 |
|
Low TEX |
2 |
0.93 |
1.03 |
0.93 |
1.18 |
|
High TEX |
1 |
0.97 |
1.08 |
1.08 |
1.43 |
|
High TEX |
2 |
0.91 |
1.04 |
0.96 |
1.19 |
The results of Lab 4 differ from the results of the other laboratories. Lab 4 did not use a large volume injection system, resulting in a too low signal.
The results of the other three laboratories show that modification of ISO 9377-2 by extending the range of hydrocarbons to C7 – C40 and excluding the aromatic hydrocarbons toluene, ethyl benzene and the isomers of xylene, have little or no effect compared to ISO 9377-2.
