Council Regulation (EC) No 440/2008 of 30 May 2008 laying down test methods pursuant to Regulation (EC) No 1907/2006 of the European Parliament and of the Council on the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) (Text with EEA relevance)

3. REFERENCES

• (1)DIN 55672(1995) Gelpermeationschromatographie (GPC) mit Tetrahydrofuran (THF) als Elutionsmittel, Teil 1.

• (2)Yau, W.W., Kirkland, J.J., and Bly, D.D. eds., (1979) Modern Size Exclusion Liquid Chromatography, J. Wiley and Sons.

• (3)ASTM D 3536-91, (1991). Standard Test Method for Molecular Weight Averages and Molecular Weight Distribution by Liquid Exclusion Chromatography (Gel Permeation Chromatography-GPC) American Society for Testing and Materials, Philadelphia, Pennsylvania.

• (4)ASTM D 5296-92, (1992) Standard Test Method for Molecular Weight Averages and Molecular Weight Distribution of Polystyrene by High Performance Size-Exclusion Chromatography. American Society for Testing and Materials, Philadelphia, Pennsylvania.

Appendix

Examples of other methods for determination of number average molecular weight (Mn) for polymers

Gel permeation chromatography (GPC) is the preferred method for determination of Mn, especially when a set of standards are available, whose structure are comparable with the polymer structure. However, where there are practical difficulties in using GPC or there is already an expectation that the substance will fail a regulatory Mn criterion (and which needs confirming), alternative methods are available, such as:

1. Use of colligative properties

• 1.1.Ebullioscopy/Cryoscopy

• involves measurement of boiling point elevation (ebullioscopy) or freezing point depression (cryoscopy) of a solvent, when the polymer is added. The method relies on the fact that the effect of the dissolved polymer on the boiling/freezing point of the liquid is dependent on the molecular weight of the polymer (1) (2).

• Applicability, Mn < 20 000 .

• 1.2.Lowering of vapour pressure

• involves the measurement of the vapour pressure of a chosen reference liquid before and after the addition of known quantities of polymer (1) (2).

• Applicability, Mn < 20 000 (theoretically; in practice however of limited value).

• 1.3Membrane osmometry

• relies on the principle of osmosis, i.e. the natural tendency of solvent molecules to pass through a semi-permeable membrane from a dilute to a concentrated solution to achieve equilibrium. In the test, the dilute solution is at zero concentration, whereas the concentrated solution contains the polymer. The effect of drawing solvent through the membrane causes a pressure differential that is dependent on the concentration and the molecular weight of the polymer (1) (3) (4).

• Applicability, Mn between 20 000 - 200 000 .

• 1.4Vapour phase osmometry

• involves comparison of the rate of evaporation of a pure solvent aerosol to at least three aerosols containing the polymer at different concentrations (1)(2)(4).

• Applicability, Mn < 20 000 .

2. End-group analysis

To use this method, knowledge of both the overall structure of the polymer and the nature of the chain terminating end groups is needed (which must be distinguishable from the main skeleton by, e.g. NMR or titration/derivatisation). The determination of the molecular concentration of the end groups present on the polymer can lead to a value for the molecular weight (7) (8) (9).

Applicability, Mn up to 50 000 (with decreasing reliability).

3. References

• (1)Billmeyer, F.W. Jr., (1984) Textbook of Polymer Science, 3rd Edn., John Wiley, New York.

• (2)Glover, C.A., (1975) Absolute Colligative Property Methods. Chapter 4. In: Polymer Molecular Weights, Part I P.E. Slade, Jr. ed., Marcel Dekker, New York.

• (3)ASTM D 3750-79, (1979) Standard Practice for Determination of Number-Average Molecular Weight of Polymers by Membrane Osmometry. American Society for Testing and Materials, Philadelphia, Pennsylvania.

• (4)Coll, H. (1989) Membrane Osmometry. In: Determination of Molecular Weight, A.R. Cooper ed., J. Wiley and Sons, pp. 25-52.

• (5)ASTM 3592-77, (1977) Standard Recommended Practice for Determination of Molecular Weight by Vapour Pressure, American Society for Testing and Materials, Philadelphia, Pennsylvania.

• (6)Morris, C.E.M., (1989) Vapour Pressure Osmometry. In: Determinationn of Molecular Weight, A.R. Cooper ed., John Wiley and Sons.

• (7)Schröder, E., Müller, G., and Arndt, K-F., (1989) Polymer Characterisation, Carl Hanser Verlag, Munich.

• (8)Garmon, R.G., (1975) End-Group Determinations, Chapter 3 In: Polymer Molecular Weights, Part I, P.E. Slade, Jr. ed., Marcel Dekker, New York.

• (9)Amiya, S., et al. (1990) Pure and Applied Chemistry, 62, 2139-2146.

A.19.   LOW MOLECULAR WEIGHT CONTENT OF POLYMERS

1. METHOD

This Gel Permeation Chromatographic method is a replicate of the OECD TG 119 (1996). The fundamental principles and further technical information are given in the references.

1.1.   INTRODUCTION

Since the properties of polymers are so varied, it is impossible to describe one single method setting out precisely the conditions for separation and evaluation which cover all eventualities and specificities occurring in the separation of polymers. In particular, complex polymer systems are often not amenable to gel permeation chromatography (GPC). When GPC is not practicable, the molecular weight may be determined by means of other methods (see Appendix). In such cases, full details and justification should be given for the method used.

The method described is based on DIN Standard 55672 (1). Detailed information about how to carry out the experiments and how to evaluate the data can be found in this DIN Standard. In case modifications of the experimental conditions are necessary, these changes must be justified. Other standards may be used, if fully referenced. The method described uses polystyrene samples of known polydispersity for calibration and it may have to be modified to be suitable for certain polymers, e.g. water soluble and long-chain branched polymers.

1.2.   DEFINITIONS AND UNITS

Low molecular weight is arbitrarily defined as a molecular weight below 1 000 dalton.

The number-average molecular weight Mn and the weight average molecular weight Mw are determined using the following equations:

Image omitted; see official source	Image omitted; see official source

where,

Hi	=	the level of the detector signal from the baseline for the retention volume Vi,
---	---	---
Mi	=	the molecular weight of the polymer fraction at the retention volume Vi, and n is the number of data points

The breadth of the molecular weight distribution, which is a measure of the dispersity of the system, is given by the ratio Mw/Mn.

1.3.   REFERENCE SUBSTANCES

Since GPC is a relative method, calibration must be undertaken. Narrowly distributed, linearly constructed polystyrene standards with known average molecular weights Mn and Mw and a known molecular weight distribution are normally used for this. The calibration curve can only be used in the determination of the molecular weight of the unknown sample if the conditions for the separation of the sample and the standards have been selected in an identical manner.

A determined relationship between the molecular weight and elution volume is only valid under the specific conditions of the particular experiment. The conditions include, above all, the temperature, the solvent (or solvent mixture), the chromatography conditions and the separation column or system of columns.

The molecular weights of the sample determined in this way are relative values and are described as ‘polystyrene equivalent molecular weights’. This means that dependent on the structural and chemical differences between the sample and the standards, the molecular weights can deviate from the absolute values to a greater or a lesser degree. If other standards are used, e.g. polyethylene glycol, polyethylene oxide, polymethyl methacrylate, polyacrylic acid, the reason should be stated.

1.4.   PRINCIPLE OF THE TEST METHOD

Both the molecular weight distribution of the sample and the average molecular weights (Mn, Mw) can be determined using GPC. GPC is a special type of liquid chromatography in which the sample is separated according to the hydrodynamic volumes of the individual constituents (2).

Separation is effected as the sample passes through a column which is filled with a porous material, typically an organic gel. Small molecules can penetrate the pores whereas large molecules are excluded. The path of the large molecules is thereby shorter and these are eluted first. The medium-sized molecules penetrate some of the pores and are eluted later. The smallest molecules, with a mean hydrodynamic radius smaller than the pores of the gel, can penetrate all of the pores. These are eluted last.

In an ideal situation, the separation is governed entirely by the size of the molecular species, but in practice it is difficult to avoid at least some absorption effects interfering. Uneven column packing and dead volumes can worsen the situation (2).

Detection is effected by e.g. refractive index or UV-absorption and yields a simple distribution curve. However, to attribute actual molecular weight values to the curve, it is necessary to calibrate the column by passing down polymers of known molecular weight and, ideally, of broadly similar structure, e.g. various polystyrene standards. Typically a Gaussian curve results, sometimes distorted by a small tail to the low molecular weight side, the vertical axis indicating the quantity, by weight, of the various molecular weight species eluted, and the horizontal axis the log molecular weight.

The low molecular weight content is derived from this curve. The calculation can only be accurate if the low molecular weight species respond equivalently on a per mass basis to the polymer as a whole.

1.5.   QUALITY CRITERIA

The repeatability (Relative Standard Deviation: RSD) of the elution volume should be better than 0,3  %. The required repeatability of the analysis has to be ensured by correction via an internal standard if a chromatogram is evaluated time-dependently and does not correspond to the above mentioned criterion (1). The polydispersities are dependent on the molecular weights of the standards. In the case of polystyrene standards typical values are:

Mp < 2 000	Mw/Mn < 1,20
2 000 < Mp < 106	Mw/Mn < 1,05
Mp > 106	Mw/Mn < 1,20

(Mp is the molecular weight of the standard at the peak maximum)

1.6.   DESCRIPTION OF THE TEST METHOD

1.6.1.   Preparation of the standard polystyrene solutions

The polystyrene standards are dissolved by careful mixing in the chosen eluent. The recommendations of the manufacturer must be taken into account in the preparation of the solutions.

The concentrations of the standards chosen are dependent on various factors, e.g. injection volume, viscosity of the solution and sensitivity of the analytical detector. The maximum injection volume must be adapted to the length of the column, in order to avoid overloading. Typical injection volumes for analytical separations using GPC with a column of 30 cm × 7,8  mm are normally between 40 and 100 μl. Higher volumes are possible, but they should not exceed 250 μl. The optimal ratio between the injection volume and the concentration must be determined prior to the actual calibration of the column.

1.6.2.   Preparation of the sample solution

In principle, the same requirements apply to the preparation of the sample solutions. The sample is dissolved in a suitable solvent, e.g. tetrahydrofuran (THF), by shaking carefully. Under no circumstances should it be dissolved using an ultrasonic bath. When necessary, the sample solution is purified via a membrane filter with a pore size of between 0,2 and 2 μm.

The presence of undissolved particles must be recorded in the final report as these may be due to high molecular weight species. An appropriate method should be used to determine the percentage by weight of the undissolved particles. The solutions should be used within 24 hours.

1.6.3.   Correction for content of impurities and additives

Correction of the content of species of M < 1 000 for the contribution from non-polymer specific components present (e.g. impurities and/or additives) is usually necessary, unless the measured content is already < 1 %. This is achieved by direct analysis of the polymer solution or the GPC eluate.

In cases where the eluate, after passage through the column, is too dilute for a further analysis it must be concentrated. It may be necessary to evaporate the eluate to dryness and dissolve it again. Concentration of the eluate must be effected under conditions which ensure that no changes occur in the eluate. The treatment of the eluate after the GPC step is dependent on the analytical method used for the quantitative determination.

1.6.4.   Apparatus

GPC apparatus comprises the following components:

• — solvent reservoir,

• — degasser (where appropriate),

• — pump,

• — pulse dampener (where appropriate),

• — injection system,

• — chromatography columns,

• — detector,

• — flowmeter (where appropriate),

• — data recorder-processor,

• — waste vessel.

It must be ensured that the GPC system is inert with regard to the utilised solvents (e.g. by the use of steel capillaries for THF solvent).

1.6.5.   Injection and solvent delivery system

A defined volume of the sample solution is loaded onto the column either using an auto-sampler or manually in a sharply defined zone. Withdrawing or depressing the plunger of the syringe too quickly, if done manually, can cause changes in the observed molecular weight distribution. The solvent-delivery system should, as far as possible, be pulsation-free ideally incorporating a pulse dampener. The flow rate is of the order of 1 ml/min.

1.6.6.   Column

Depending on the sample, the polymer is characterised using either a simple column or several columns connected in sequence. A number of porous column materials with defined properties (e.g. pore size, exclusion limits) are commercially available. Selection of the separation gel or the length of the column is dependent on both the properties of the sample (hydrodynamic volumes, molecular weight distribution) and the specific conditions for separation such as solvent, temperature and flow rate (1) (2) (3).

1.6.7.   Theoretical plates

The column or the combination of columns used for separation must be characterised by the number of theoretical plates. This involves, in the case of THF as elution solvent, loading a solution of ethyl benzene or other suitable non-polar solute onto a column of known length. The number of theoretical plates is given by the following equation:

Image omitted; see official source	or	Image omitted; see official source

where,

N	=	the number of theoretical plates
---	---	---
Ve	=	the elution volume at the peak maximum
---	---	---
W	=	the baseline peak width
---	---	---
W1/2	=	the peak width at half height

1.6.8.   Separation efficiency

In addition to the number of theoretical plates, which is a quantity determining the bandwidth, a part is also played by the separation efficiency, this being determined by the steepness of the calibration curve. The separation efficiency of a column is obtained from the following relationship:

Image omitted; see official source

where,

Ve, Mx	=	the elution volume for polystyrene with the molecular weight Mx
---	---	---
Ve,(10.Mx)	=	the elution volume for polystyrene with a ten times greater molecular weight

The resolution of the system is commonly defined as follows:

Image omitted; see official source

where,

Ve1, Ve2	=	the elution volumes of the two polystyrene standards at the peak maximum
---	---	---
W1, W2	=	the peak widths at the base-1ine
---	---	---
M1, M2	=	the molecular weights at the peak maximum (should differ by a factor of 10).

The R-value for the column system should be greater than 1,7 (4).

1.6.9.   Solvents

All solvents must be of high purity (for THF purity of 99,5  % is used). The solvent reservoir (if necessary in an inert gas atmosphere) must be sufficiently large for the calibration of the column and several sample analyses. The solvent must be degassed before it is transported to the column via the pump.

1.6.10.   Temperature control

The temperature of the critical internal components (injection loop, columns, detector and tubing) should be constant and consistent with the choice of solvent.

1.6.11.   Detector

The purpose of the detector is to record quantitatively the concentration of sample eluted from the column. In order to avoid unnecessary broadening of peaks the cuvette volume of the detector cell must be kept as small as possible. It should not be larger than 10 μl except for light scattering and viscosity detectors. Differential refractometry is usually used for detection. However, if required by the specific properties of the sample or the elution solvent, other types of detectors can be used, e.g. UV/VIS, IR, viscosity detectors, etc.

2. DATA AND REPORTING

2.1.   DATA

The DIN Standard (1) should be referred to for the detailed evaluation criteria as well as for the requirements relating to the collecting and processing of data.

For each sample, two independent experiments must be carried out. They have to be analysed individually. In all cases it is essential to determine also data from blanks, treated under the same conditions as the sample.

It is necessary to indicate explicitly that the measured values are relative values equivalent to the molecular weights of the standard used.

After determination of the retention volumes or the retention times (possibly corrected using an internal standard), log Mp values (Mp being the peak maxima of the calibration standard) are plotted against one of those quantities. At least two calibration points are necessary per molecular weight decade, and at least five measurement points are required for the total curve, which should cover the estimated molecular weight of the sample. The low molecular weight end-point of the calibration curve is defined by n-hexyl benzene or another suitable non-polar solute. The portion of the curve corresponding to molecular weights below 1 000 is determined and corrected as necessary for impurities and additives. The elution curves are generally evaluated by means of electronic data processing. In case manual digitisation is used, ASTM D 3536-91 can be consulted (3).

If any insoluble polymer is retained on the column, its molecular weight is likely to be higher than that of the soluble fraction, and if not considered would result in an overestimation of the low molecular weight content. Guidance for correcting the low molecular weight content for insoluble polymer is provided in the Appendix.

The distribution curve must be provided in the form of a table or as figure (differential frequency or sum percentages against log M). In the graphic representation, one molecular weight decade should be normally about 4 cm in width and the peak maximum should be about 8 cm in height. In the case of integral distribution curves the difference in the ordinate between 0 and 100 % should be about 10 cm.

2.2.   TEST REPORT

The test report must include the following information:

2.2.1.   Test substance:

• — available information about test substance (identity, additives, impurities),

• — description of the treatment of the sample, observations, problems.

2.2.2.   Instrumentation:

• — reservoir of eluent, inert gas, degassing of the eluent, composition of the eluent, impurities,

• — pump, pulse dampener, injection system,

• — separation columns (manufacturer, all information about the characteristics of the columns, such as pore size, kind of separation material, etc., number, length and order of the columns used),

• — number of the theoretical plates of the column (or combination), separation efficiency (resolution of the system),

• — information on symmetry of the peaks,

• — column temperature, kind of temperature control,

• — detector (measurement principle, type, cuvette volume),

• — flowmeter if used (manufacturer, measurement principle),

• — system to record and process data (hardware and software).

2.2.3.   Calibration of the system:

• — detailed description of the method used to construct the calibration curve,

• — information about quality criteria for this method (e.g. correlation coefficient, error sum of squares, etc.),

• — information about all extrapolations, assumptions and approximations made during the experimental procedure and the evaluation and processing of data,

• — all measurements used for constructing the calibration curve have to be documented in a table which includes the following information for each calibration point: — name of the sample, — manufacturer of the sample, — characteristic values of the standards Mp, Mn, Mw, Mw/Mn, as provided by the manufacturer or derived by subsequent measurements, together with details about the method of determination, — injection volume and injection concentration, — Mp value used for calibration, — elution volume or corrected retention time measured at the peak maxima, — Mp calculated at the peak maximum, — percentage error of the calculated Mp and the calibration value.

2.2.4.   Information on the low molecular weight polymer content:

• — description of the methods used in the analysis and the way in which the experiments were conducted,

• — information about the percentage of the low molecular weight species content (w/w) related to the total sample,

• — information about impurities, additives and other non-polymer species in percentage by weight related to the total sample.

2.2.5.   Evaluation:

• — evaluation on a time basis: all methods to ensure the required reproducibility (method of correction, internal standard etc.),

• — information about whether the evaluation was effected on the basis of the elution volume or the retention time,

• — information about the limits of the evaluation if a peak is not completely analysed,

• — description of smoothing methods, if used,

• — preparation and pre-treatment procedures of the sample,

• — the presence of undissolved particles, if any,

• — injection volume (μl) and injection concentration (mg/ml),

• — observations indicating effects which lead to deviations from the ideal GPC profile,

• — detailed description of all modifications in the testing procedures,

• — details of the error ranges,

• — any other information and observations relevant for the interpretation of the results.

3. REFERENCES

• (1)DIN 55672 (1995) Gelpermeationschromatographie (GPC) mit Tetrahydrofuran (THF) als Elutionsmittel, Teil 1.

• (2)Yau, W.W., Kirkland, J.J., and Bly, D.D. eds. (1979) Modern Size Exclusion Liquid Chromatography, J. Wiley and Sons.

• (3)ASTM D 3536-91, (1991) Standard Test method for Molecular Weight Averages and Molecular Weight Distribution by Liquid Exclusion Chromatography (Gel Permeation Chromatography-GPC). American Society for Testing and Materials, Philadelphia, Pennsylvania.

• (4)ASTM D 5296-92, (1992) Standard Test method for Molecular Weight Averages and Molecular Weight Distribution of Polystyrene by High Performance Size-Exclusion Chromatography. American Society for Testing and Materials, Philadelphia, Pennsylvania.

Appendix

Guidance for correcting low molecular content for the presence of insoluble polymer

When insoluble polymer is present in a sample, it results in mass loss during the GPC analysis. The insoluble polymer is irreversibly retained on the column or sample filter while the soluble portion of the sample passes through the column. In the case where the refractive index increment (dn/dc) of the polymer can be estimated or measured, one can estimate the sample mass lost on the column. In that case, one makes a correction using an external calibration with standard materials of known concentration and dn/dc to calibrate the response of the refractometer. In the example hereafter a poly(methyl methacrylate) (pMMA) standard is used.

In the external calibration for analysis of acrylic polymers, a pMMA standard of known concentration in tetrahydrofuran, is analysed by GPC and the resulting data are used to find the refractometer constant according to the equation:

K = R/(C × V × dn/dc)

where:

K	=	the refractometer constant (in microvolt second/ml),
---	---	---
R	=	the response of the pMMA standard (in microvolt/second),
---	---	---
C	=	the concentration of the pMMA standard (in mg/ml),
---	---	---
V	=	the injection volume (in ml), and
---	---	---
dn/dc	=	the refractive index increment for pMMA in tetrahydrofuran (in ml/mg).

The following data are typical for a pMMA standard:

R	=	2 937 891
---	---	---
C	=	1,07  mg/ml
---	---	---
V	=	0,1  ml
---	---	---
dn/dc	=	9 × 10-5 ml/mg

The resulting K value, 3,05  × 10¹¹ is then used to calculate the theoretical detector response if 100 % of the polymer injected had eluted through the detector.

A.20.   SOLUTION/EXTRACTION BEHAVIOUR OF POLYMERS IN WATER

The full description of this test method has been deleted. The equivalent international test method, or other applicable test methods for the endpoint in question, appear in Table 1 of Part 0.

A.21.   OXIDISING PROPERTIES (LIQUIDS)

The full description of this test method has been deleted. The equivalent international test method, or other applicable test methods for the endpoint in question, appear in Table 1 of Part 0.

A.22.   LENGTH WEIGHTED GEOMETRIC MEAN DIAMETER OF FIBRES

1. METHOD

1.1.   INTRODUCTION

This method describes a procedure to measure the Length Weighted Geometric Mean Diameter (LWGMD) of bulk Man Made Mineral Fibres (MMMF). As the LWGMD of the population will have a 95 % probability of being between the 95 % confidence levels (LWGMD ± two standard errors) of the sample, the value reported (the test value) will be the lower 95 % confidence limit of the sample (i.e. LWGMD — 2 standard errors). The method is based on an update (June 1994) of a draft HSE industry procedure agreed at a meeting between ECFIA and HSE at Chester on 26/9/93 and developed for and from a second inter-laboratory trial (1, 2). This measurement method can be used to characterise the fibre diameter of bulk substances or products containing MMMFs including refractory ceramic fibres (RCF), man-made vitreous fibres (MMVF), crystalline and polycrystalline fibres.

Length weighting is a means of compensating for the effect on the diameter distribution caused by the breakage of long fibres when sampling or handling the material. Geometric statistics (geometric mean) are used to measure the size distribution of MMMF diameters because these diameters usually have size distributions that approximate to log normal.

Measuring length as well as diameter is both tedious and time consuming but, if only those fibres that touch an infinitely thin line on a SEM field of view are measured, then the probability of selecting a given fibre is proportional to its length. As this takes care of the length in the length weighting calculations, the only measurement required is the diameter and the LWGMD-2SE can be calculated as described.

1.2.   DEFINITIONS

Particle: An object with a length to width ratio of less than 3:1.

Fibre: An object with a length to with ratio (aspect ratio) of at least 3:1.

1.3.   SCOPE AND LIMITATIONS

The method is designed to look at diameter distributions which have median diameters from 0,5 μm to 6 μm. Larger diameters can be measured by using lower SEM magnifications but the method will be increasingly limited for finer fibre distributions and a TEM (transmission electron microscope) measurement is recommended if the median diameter is below 0,5 μm.

1.4.   PRINCIPLE OF THE TEST METHOD

A number of representative core samples are taken from the fibre blanket or from loose bulk fibre. The bulk fibres are reduced in length using a crushing procedure and a representative sub-sample dispersed in water. Aliquots are extracted and filtered through a 0,2 μm pore size, polycarbonate filter and prepared for examination using scanning electron microscope (SEM) techniques. The fibre diameters are measured at a screen magnification of × 10 000 or greater ([^E0007]) using a line intercept method to give an unbiased estimate of the median diameter. The lower 95 % confidence interval (based on a one sided test) is calculated to give an estimate of the lowest value of the geometric mean fibre diameter of the material.

1.5.   DESCRIPTION OF THE TEST METHOD

1.5.1.   Safety/precautions

Personal exposure to airborne fibres should be minimised and a fume cupboard or glove box should be used for handling the dry fibres. Periodic personal exposure monitoring should be carried out to determine the effectiveness of the control methods. When handling MMMF’s disposable gloves should be worn to reduce skin irritation and to prevent cross-contamination.

1.5.2.   Apparatus/equipment

• — Press and dyes (capable of producing 10 MPa).

• — 0,2 μm pore size polycarbonate capillary pore filters (25 mm diameter).

• — 5 μm pore size cellulose ester membrane filter for use as a backing filter.

• — Glass filtration apparatus (or disposable filtration systems) to take 25 mm diameter filters (e.g. Millipore glass microanalysis kit, type No XX10 025 00).

• — Freshly distilled water that has been filtered through a 0,2 μm pore size filter to remove micro-organisms.

• — Sputter coater with a gold or gold/palladium target.

• — Scanning electron microscope capable of resolving down to 10 nm and operating at × 10 000 magnification.

• — Miscellaneous: spatulas, type 24 scalpel blade, tweezers, SEM tubes, carbon glue or carbon adhesive tape, silver dag.

• — Ultrasonic probe or bench top ultrasonic bath.

• — Core sampler or cork borer, for taking core samples from MMMF blanket.

1.5.3.   Test Procedure

1.5.3.1.   Sampling

For blankets and bats a 25 mm core sampler or cork borer is used to take samples of the cross-section. These should be equally spaced across the width of a small length of the blanket or taken from random areas if long lengths of the blanket are available. The same equipment can be used to extract random samples from loose fibre. Six samples should be taken when possible, to reflect spatial variations in the bulk material.

The six core samples should be crushed in a 50 mm diameter dye at 10 MPa. The material is mixed with spatula and re-pressed at 10 MPa. The material is then removed from the dye and stored in a sealed glass bottle.

1.5.3.2.   Sample Preparation

If necessary, organic binder can be removed by placing the fibre inside a furnace at 450 °C for about one hour.

Cone and quarter to subdivide the sample (this should be done inside a dust cupboard).

Using a spatula, add a small amount (< 0,5 g) of sample to 100 ml of freshly distilled water that has been filtered through a 0,2 μm membrane filter (alternative sources of ultra pure water may be used if they are shown to be satisfactory). Disperse thoroughly by the use of an ultrasonic probe operated at 100 W power and tuned so that cavitation occurs. (If a probe is not available use the following method: repeatedly shake and invert for 30 seconds; ultrasonic in a bench top ultrasonic bath for five minutes; then repeatedly shake and invert for a further 30 seconds.)

Immediately after dispersion of the fibre, remove a number of aliquots (e.g. three aliquots of 3, 6 and 10 ml) using a wide-mouthed pipette (2-5 ml capacity).

Vacuum filter each aliquot through a 0,2 μm polycarbonate filter supported by a 5 μm pore MEC backing filter, using a 25 mm glass filter funnel with a cylindrical reservoir. Approximately 5 ml of filtered distilled water should be placed into the funnel and the aliquot slowly pipetted into the water holding the pipette tip below the meniscus. The pipette and the reservoir must be flushed thoroughly after pipetting, as thin fibres have a tendency to be located more on the surface.

Carefully remove the filter and separate it from the backing filter before placing it in a container to dry.

Cut a quarter or half filter section of the filtered deposit with a type 24 scalpel blade using a rocking action. Carefully attach the cut section to a SEM stub using a sticky carbon tab or carbon glue. Silver dag should be applied in at least three positions to improve the electrical contact at the edges of the filter and the stub. When the glue/silver dag is dry, sputter coat approximately 50 nm of gold or gold/palladium onto the surface of the deposit.

1.5.3.3.   SEM calibration and operation

1.5.3.3.1.   Calibration

The SEM calibration should be checked at least once a week (ideally once a day) using a certified calibration grid. The calibration should be checked against a certified standard and if the measured value (SEM) is not within ± 2  % of the certified value, then the SEM calibration must be adjusted and re-checked.

The SEM should be capable of resolving at least a minimum visible diameter of 0,2 μm, using a real sample matrix, at a magnification of × 2 000 .

1.5.3.3.2.   Operation

The SEM should be operated at 10 000 magnification ([^E0008]) using conditions that give good resolution with an acceptable image at slow scan rates of, for example, 5 seconds per frame. Although the operational requirements of different SEMs may vary, generally to obtain the best visibility and resolution, with relatively low atomic weight materials, accelerating voltages of 5-10 keV should be used with a small spot size setting and short working distance. As a linear traverse is being conducted, then a 0° tilt should be used to minimise re-focussing or, if the SEM has a eucentric stage, the eucentric working distance should be used. Lower magnification may be used if the material does not contain small (diameter) fibres and the fibre diameters are large (> 5 μm).

1.5.3.4.   Sizing

1.5.3.4.1.   Low magnification examination to assess the sample

Initially the sample should be examined at low magnification to look for evidence of clumping of large fibres and to assess the fibre density. In the event of excessive clumping it is recommended that a new sample is prepared.

For statistical accuracy it is necessary to measure a minimum number of fibres and high fibre density may seem desirable as examining empty fields is time consuming and does not contribute to the analysis. However, if the filter is overloaded, it becomes difficult to measure all the measurable fibres and, because small fibres may be obscured by larger ones, they may be missed.

Bias towards over estimating the LWGMD may result from fibre densities in excess of 150 fibres per millimetre of linear traverse. On the other hand, low fibre concentrations will increase the time of analysis and it is often cost effective to prepare a sample with a fibre density closer to the optimum than to persist with counts on low concentration filters. The optimum fibre density should give an average of about one or two countable fibre per fields of view at 5 000 magnification. Nevertheless the optimum density will depend on the size (diameter) of the fibres, so it is necessary that the operator uses some expert judgement in order to decide whether the fibre density is close to optimal or not.

1.5.3.4.2.   Length weighting of the fibre diameters

Only those fibres that touch (or cross) an (infinitely) thin line drawn on the screen of the SEM are counted. For this reason a horizontal (or vertical) line is drawn across the centre of the screen.

Alternatively a single point is placed at the centre of the screen and a continuous scan in one direction across the filter is initiated. Each fibre of aspect ratio grater than 3:1 touching or crossing this point has its diameter measured and recorded.

1.5.3.4.3.   Fibre sizing

It is recommended that a minimum of 300 fibres are measured. Each fibre is measured only once at the point of intersection with the line or point drawn on the image (or close to the point of intersection if the fibre edges are obscured). If fibres with non-uniform cross sections are encountered, a measurement representing the average diameter of the fibre should be used. Care should be taken in defining the edge and measuring the shortest distance between the fibre edges. Sizing may be done on line, or off-line on stored images or photographs. Semi-automated image measurement systems that download data directly into a spreadsheet are recommended, as they save time, eliminate transcription errors and calculations can be automated.

The ends of long fibres should be checked at low magnification to ensure that they do not curl back into the measurement field of view and are only measured once.

2. DATA

2.1.   TREATMENT OF RESULTS

Fibre diameters do not usually have a normal distribution. However, by performing a log transformation it is possible to obtain a distribution that approximates to normal.

Calculate the arithmetic mean (mean lnD) and the standard deviation (SDlnD) of the log to base e values (lnD) of the n fibre diameters (D).

Image omitted; see official source	(1)
Image omitted; see official source	(2)

The standard deviation is divided by the square root of the number of measurements (n) to obtain the standard error (SElnD).

Image omitted; see official source	(3)

Subtract two times the standard error from the mean and calculate the exponential of this value (mean minus two standard errors) to give the geometric mean minus two geometric standard errors.

Image omitted; see official source	(4)

3. REPORTING

TEST REPORT

The test report should include at least the following information:

• — The value of LWGMD-2SE.

• — Any deviations and particularly those which may have an effect on the precision or accuracy of the results with appropriate justifications.

4. REFERENCES

• 1. B. Tylee SOP MF 240. Health and Safety Executive, February 1999.

• 1. G. Burdett and G. Revell. Development of a standard method to measure the length-weigthed geometric mean fibre diameter: Results of the Second inter-laboratory exchange. IR/L/MF/94/07. Project R42.75 HPD. Health and Safety Executive, Research and Laboratory Services Division, 1994.

A.23.   PARTITION COEFFICIENT (1-OCTANOL/WATER): SLOW-STIRRING METHOD

INTRODUCTION

• 1.This Test Method is equivalent to OECD Test Guideline (TG) 123 (2006). 1-octanol/water partition coefficient (POW) values up to a log POW of 8,2 have been accurately determined by the slow-stirring method (1). Therefore it is a suitable experimental approach for the direct determination of POW of highly hydrophobic substances.

• 2.Other methods for the determination of the 1-octanol/water partition coefficient (POW) are the ‘shake-flask’ method (2), and the determination of the POW from reversed phase HPLC-retention behaviour (3). The ‘shake-flask’ method is prone to artifacts due to transfer of octanol micro-droplets into the aqueous phase. With increasing values of POW the presence of these droplets in the aqueous phase leads to an increasing overestimation of the concentration of the test substance in the water. Therefore, its use is limited to substances with log POW < 4. The second method relies on solid data of directly determined POW values to calibrate the relationship between HPLC-retention behaviour and measured values of POW. A draft OECD guideline was available for determining 1-octanol/water partition coefficients of ionisable substances (4) but shall no longer be used.

• 3.This Test Method has been developed in The Netherlands. The precision of the methods described here has been validated and optimized in a ring-test validation study in which 15 laboratories participated (5).

INITIAL CONSIDERATIONS

Significance and use

• 4.For inert organic substances highly significant relationships have been found between 1-octanol/water partition coefficients (POW) and their bioaccumulation in fish. Moreover, POW has been demonstrated to be correlated to fish toxicity as well as to sorption of chemicals to solids such as soils and sediments. An extensive overview of the relationships has been given in reference (6).

• 5.A wide variety of relationships between the 1-octanol/water partition coefficient and other substance properties of relevance to environmental toxicology and chemistry have been established. As a consequence, the 1-octanol/water partition coefficient has evolved as a key parameter in the assessment of the environmental risk of chemicals as well as in the prediction of fate of chemicals in the environment.

Scope

• 6.The slow-stirring experiment is thought to reduce the formation of micro-droplets from 1-octanol droplets in the water phase. As a consequence, overestimation of the aqueous concentration due to test substance molecules associated to such droplets does not occur. Therefore, the slow-stirring method is particularly suitable for the determination of POW for substances with expected log POW values of 5 and higher, for which the shake-flask method (2) is prone to yield erroneous results.

DEFINITION AND UNITS

• 7.The partition coefficient of a substance between water and a lipophilic solvent (1-octanol) characterizes the equilibrium distribution of the chemical between the two phases. The partition coefficient between water and 1-octanol (POW) is defined as the ratio of the equilibrium concentrations of the test substance in 1-octanol saturated with water (CO) and water saturated with 1-octanol (CW).

Image omitted; see official source

• As a ratio of concentrations it is dimensionless. Most frequently it is given as the logarithm to the base 10 (log POW). POW is temperature dependent and reported data should include the temperature of the measurement.

PRINCIPLE OF THE METHOD

• 8.In order to determine the partitioning coefficient, water, 1-octanol, and the test substance are equilibrated with each other at constant temperature. Then the concentrations of the test substance in the two phases are determined.

• 9.The experimental difficulties associated with the formation of micro-droplets during the shake-flask experiment can be reduced in the slow-stirring experiment proposed here. In the slow-stirring experiment, water, 1-octanol and the test substance are equilibrated in a thermostated stirred reactor. Exchange between the phases is accelerated by stirring. The stirring introduces limited turbulence which enhances the exchange between 1-octanol and water without micro-droplets being formed (1).

APPLICABILITY OF THE TEST

• 10.Since the presence of substances other than the test substance might influence the activity coefficient of the test substance, the test substance should be tested as a pure substance. The highest purity commercially available should be employed for the 1-octanol/water partition experiment.

• 11.The present method applies to pure substances that do not dissociate or associate and that do not display significant interfacial activity. It can be applied to determine the 1-octanol/water partition ratio of such substances and of mixtures. When the method is used for mixtures, the 1-octanol/water partition ratios determined are conditional and depend on the chemical composition of the mixture tested and on the electrolyte composition employed as aqueous phase. Provided additional steps are taken, the method is also applicable to dissociating or associating compounds (paragraph 12).

• 12.Due to the multiple equilibria in water and 1-octanol involved in the 1-octanol/water partitioning of dissociating substances such as organic acids and phenols, organic bases, and organometallic substances, the 1-octanol/water partition ratio is a conditional constant strongly dependent on electrolyte composition (7)(8). Determination of the 1-octanol/water partition ratio therefore requires that pH and electrolyte composition be controlled in the experiment and reported. Expert judgement has to be employed in the evaluation of these partition ratios. Using the value of dissociation constant(s), suitable pH-values need to be selected, such that a partitioning ratio is determined for each ionization state. Non-complexing buffers must be used when testing organometallic compounds (8). Taking the existing knowledge on the aqueous chemistry (complexation constants, dissociation constants) into account, the experimental conditions should be chosen in such a manner that the speciation of the test substance in the aqueous phase can be estimated. The ionic strength should be identical in all experiments by employing a background electrolyte.

• 13.Difficulties in the test may arise in conducting the test for substances with low water solubility or high POW, due to the fact that the concentrations in the water become very low such that their accurate determination is difficult. This Test Method provides guidance on how to deal with this problem.

INFORMATION ON THE TEST SUBSTANCE

• 14.Chemical reagents should be of analytical grade or of higher purity. The use of non-labelled test substances with known chemical composition and preferably at least 99 % purity, or of radiolabelled test substances with known chemical composition and radiochemical purity, is recommended. In the case of short half-life tracers, decay corrections should be applied. In the case of radiolabelled test substances, a chemical specific analytical method should be employed to ensure that the measured radioactivity is directly related to the test substance.

• 15.An estimate of log POW may be obtained by using commercially available software for estimation of log POW, or by using the ratio of the solubilities in both solvents.

• 16.Before carrying out a slow-stirring experiment for determination of POW, the following information on the test substance should be available:

• (a) structural formula

• (b) suitable analytical methods for determination of the concentration of the substance in water and 1-octanol

• (c) dissociation constant(s) of ionisable substances (OECD Guideline 112 (9))

• (d) aqueous solubility (10)

• (e) abiotic hydrolysis (11)

• (f) ready biodegradability (12)

• (g) vapour pressure (13).

DESCRIPTION OF THE METHOD

Equipment and apparatus

• 17.Standard laboratory equipment is required, in particular, the following:

• — magnetic stirrers and Teflon coated magnetic stir bars are employed to stir the water phase;

• — analytical instrumentation, suitable for determination of the concentration of the test substance at the expected concentrations;

• — stirring-vessel with a tap at the bottom. Dependent on the estimate of log POW and the Limit of Detection (LOD) of the test compound, the use of a reaction vessel of the same geometry larger than one litre has to be considered, so that a sufficient volume of water can be obtained for chemical extraction and analysis. This will result in higher concentrations in the water extract and thus a more reliable analytical determination. A table giving estimates of the minimum volume needed, the LOD of the compound, its estimated log POW and its water solubility is given in Appendix 1. The table is based on the relationship between log POW and the ratio between the solubilities in octanol and water, as presented by Pinsuwan et al. (14): Image omitted; see official source where Image omitted; see official source (in molarity); and the relationship given by Lyman (15) for predicting water solubility. Water solubilities calculated with the equation given in Appendix 1 must be seen as a first estimate. It should be noted that the user is free to generate an estimate of water solubility by means of any relationship that is considered to better represent the relationship between hydrophobicity and solubility. For solid compounds, inclusion of melting point in the prediction of solubility is for instance recommended. In case a modified equation is used, it should be ascertained that the equation for calculation of solubility in octanol is still valid. A schematic drawing of a glass-jacketed stirring-vessel with a volume of ca. one litre is given in Appendix 2. The proportions of the vessel shown in Appendix 2 have proven favourable and should be maintained when apparatus of a different size is used;

• — a means for keeping the temperature constant during the slow-stirring experiment is essential.

• 18.Vessels should be made from inert material such that adsorption to vessel surfaces is negligible.

Preparation of the test solutions

• 19.The POW determination should be carried out with the highest purity 1-octanol that is commercially available (at least + 99 %). Purification of 1-octanol by extraction with acid, base and water and subsequent drying is recommended. In addition, distillation can be used to purify 1-octanol. Purified 1-octanol is to be used to prepare standard solutions of the test substances. Water to be used in the POW determination should be glass or quartz distilled, or obtained from a purification system, or HPLC-grade water may be used. Filtration through a 0,22 μm filter is required for distilled water, and blanks should be included to check that no impurities are in the concentrated extracts that may interfere with the test substance. If a glass fibre filter is used, it should be cleaned by baking for at least three hours at 400 °C.

• 20.Both solvents are mutually saturated prior to the experiment by equilibrating them in a sufficiently large vessel. This is accomplished by slow-stirring the two-phase system for two days.

• 21.An appropriate concentration of test substance is selected and dissolved in 1-octanol (saturated with water). The 1-octanol/water partition coefficient needs to be determined in dilute solutions in 1-octanol and water. Therefore the concentration of the test substance should not exceed 70 % of its solubility with a maximum concentration of 0,1 M in either phase (1). The 1-octanol solutions used for the experiment must be devoid of suspended solid test substance.

• 22.The appropriate amount of test substance is dissolved in 1-octanol (saturated with water). If the estimate of log POW exceeds five, care has to be taken that the 1-octanol solutions used for the experiment are devoid of suspended solid test substance. To that end, the following procedure for chemicals with an estimated value of log POW > 5 is followed:

• — the test substance is dissolved in 1-octanol (saturated with water);

• — the solution is given sufficient time for the suspended solid substance to settle out. During the settling period, the concentration of the test substance is monitored;

• — after the measured concentrations in the 1-octanol-solution have attained stable values, the stock solution is diluted with an appropriate volume of 1-octanol;

• — the concentration of the diluted stock solution is measured. If the measured concentration is consistent with the dilution, the diluted stock solution can be employed in the slow-stirring experiment.

Extraction and analysis of samples

• 23.A validated analytical method should be used for the assay of test substance. The investigators have to provide evidence that the concentrations in the water saturated 1-octanol as well as in the 1-octanol saturated water phase during the experiment are above the method limit of quantification of the analytical procedures employed. Analytical recoveries of the test substance from the water phase and from the 1-octanol phase need to be established prior to the experiment in those cases for which extraction methods are necessary. The analytical signal needs to be corrected for blanks and care should be taken that no carry-over of analyte from one sample to another can occur.

• 24.Extraction of the water phase with an organic solvent and preconcentration of extract are likely to be required prior to analysis, due to rather low concentrations of hydrophobic test substances in the water phase. For the same reason it is necessary to reduce eventual blank concentrations. To that end, it is necessary to employ high purity solvents, preferably solvents for residue analysis. Moreover, working with carefully pre-cleaned (e.g. solvent washing or baking at elevated temperature) glassware can help to avoid cross-contamination.

• 25.An estimate of log POW may be obtained from an estimation program or by expert judgment. If the value is higher than six then blank corrections and analyte carry-over need to be monitored closely. Similarly, if the estimate of log POW exceeds six, the use of a surrogate standard for recovery correction is mandatory, so that high preconcentration factors can be reached. A number of software programs for the estimation of log POW are commercially available ([^E0009]), e.g. Clog P (16), KOWWIN (17), ProLogP (18) and ACD log P (19). Descriptions of the estimation approaches can be found in references (20-22).

• 26.The limits of quantification (LOQ) for determination of the test substance in 1-octanol and water are established using accepted methods. As a rule of thumb, the method limit of quantification can be determined as the concentration in water or 1-octanol that produces a signal to noise ratio of ten. A suitable extraction and pre-concentration method should be selected and analytical recoveries should also be specified. A suitable pre-concentration factor is selected in order to obtain a signal of the required size upon analytical determination.

• 27.On the basis of the parameters of the analytical method and the expected concentrations, an approximate sample size required for an accurate determination of the compound concentration is determined. The use of water samples that are too small to obtain a sufficient analytical signal should be avoided. Also, the use of excessively large water samples should be avoided, since otherwise there might be too little water left for the minimum number of analyses required (n = 5). In Appendix 1, the minimum sample volume is indicated as a function of the vessel volume, the LOD of the test substance and the solubility of the test substance.

• 28.Quantification of the test substances occurs by comparison with calibration curves of the respective compound. The concentrations in the samples analysed must be bracketed by concentrations of standards.

• 29.For test substances with a log POW estimate higher than six a surrogate standard has to be spiked to the water sample prior to extraction in order to register losses occurring during extraction and pre-concentration of the water samples. For accurate recovery correction, the surrogates must have properties that are very close to, or identical with, those of the test substance. Preferably, (stable) isotopically-labelled analogues of the substances of interest (for example, perdeuterated or ¹³C-labelled) are used for this purpose. If the use of labelled stable isotopes, i.e. ¹³C or ²H, is not possible it should be demonstrated from reliable data in the LITERATURE that the physical-chemical properties of the surrogate are very close to those of the test substance. During liquid-liquid extraction of the water phase emulsions can form. They can be reduced by addition of salt and allowing the emulsion to settle overnight. Methods used for extracting and pre-concentrating the samples need to be reported.

• 30.Samples withdrawn from the 1-octanol phase may, if necessary, be diluted with a suitable solvent prior to analysis. Moreover, the use of surrogate standards for recovery correction is recommended for substances for which the recovery experiments demonstrated a high degree of variation in the recovery experiments (relative standard deviation > 10 %).

• 31.The details of the analytical method need to be reported. This includes the method of extraction, pre-concentration and dilution factors, instrument parameters, calibration routine, calibration range, analytical recovery of the test substance from water, addition of surrogate standards for recovery correction, blank values, detection limits and limits of quantification.

Performance of the Test

Optimal 1-octanol/water volume ratios

• 32.When choosing the water and 1-octanol volumes, the LOQ in 1-octanol and water, the pre-concentration factors applied to the water samples, the volumes sampled in 1-octanol and water, and the expected concentrations should be considered. For experimental reasons, the volume of 1-octanol in the slow-stirring system should be chosen such that the 1-octanol layer is sufficiently thick (> 0,5 cm) in order to allow for sampling of the 1-octanol phase without disturbing it.

• 33.Typical phase ratios used for the determinations of compounds with log POW of 4,5 and higher are 20 to 50 ml of 1-octanol and 950 to 980 ml of water in a one litre vessel.

Test conditions

• 34.During the test the reaction vessel is thermostated to reduce temperature variation to below 1 °C. The assay should be performed at 25 °C.

• 35.The experimental system should be protected from daylight by either performing the experiment in a dark room or by covering the reaction vessel with aluminium foil.

• 36.The experiment should be performed in a dust-free (as far as possible) environment.

• 37.The 1-octanol-water system is stirred until equilibrium is attained. In a pilot experiment the length of the equilibration period is assessed by performing a slow-stirring experiment and sampling water and 1-octanol periodically. The sampling time points should be interspersed by a minimum period of five hours.

• 38.Each POW determination has to be performed employing at least three independent slow-stirring experiments.

Determination of the equilibration time

• 39.It is assumed that the equilibrium is achieved when a regression of the 1-octanol/water concentration ratio against time over a time span of four time points yields a slope that is not significantly different from zero at a p-level of 0,05. The minimum equilibration time is one day before sampling can be started. As a rule of thumb, sampling of substances with a log POW estimate of less than five can take place during days two and three. The equilibration might have to be extended for more hydrophobic compounds. For a compound with log POW of 8,23 (decachlorobiphenyl) 144 hours were sufficient for equilibration. Equilibrium is assessed by means of repeated sampling of a single vessel.

Starting the experiment

• 40.At the start of the experiment the reaction vessel is filled with 1-octanol-saturated water. Sufficient time should be allowed to reach the thermostated temperature.

• 41.The desired amount of test substance (dissolved in the required volume of 1-octanol saturated with water) is carefully added to the reaction vessel. This is a crucial step in the experiment, since turbulent mixing of the two phases has to be avoided. To that end, the 1-octanol phase can be pipetted slowly against the wall of the experimental vessel, close to the water surface. It will subsequently flow along the glass wall and form a film above the water phase. The decantation of 1-octanol directly into the flask should always be avoided; drops of 1-octanol should not be allowed to fall directly into the water.

• 42.After starting the stirring, the stirring rate should be increased slowly. If the stirring motors cannot be appropriately adjusted the use of a transformer should be considered. The stirring rate should be adjusted so that a vortex at the interface between water and 1-octanol of 0,5 to maximally 2,5 cm depth is created. The stirring rate should be reduced if the vortex depth of 2,5 cm is exceeded; otherwise micro-droplets may be formed from 1-octanol droplets in the water phase, leading to an overestimation of the concentration of the test substance in the water. The maximum stirring rate of 2,5 cm is recommended on the basis of the findings in the ring-test validation study (5). It is a compromise between achieving a rapid rate of equilibration, while limiting the formation of 1-octanol micro-droplets.

Sampling and Sample Treatment

• 43.The stirrer should be turned off prior to sampling and the liquids should be allowed to stop moving. After sampling is completed, the stirrer is started again slowly, as described above, and then the stirring rate is increased gradually.

• 44.The water phase is sampled from a stopcock at the bottom of the reaction vessel. Always discard the dead volume of water contained in the taps (approximately 5 ml in the vessel shown in the Appendix 2). The water in the taps is not stirred and therefore not in equilibrium with the bulk. Note the volume of the water samples, and make sure that the amount of test substance present in the discarded water is taken into account when setting up a mass balance. Evaporative losses should be minimized by allowing the water to flow quiescently into the separatory funnel, such that there is no disturbance of the water/1-octanol layer.

• 45.1-Octanol samples are obtained by withdrawing a small aliquot (ca. 100 μl) from the 1-octanol layer with a 100 microlitre all glass-metal syringe. Care should be taken not to disturb the boundary. The volume of the sampled liquid is recorded. A small aliquot is sufficient, since the 1-octanol sample will be diluted.

• 46.Unnecessary sample transfer steps should be avoided. To that end the sample volume should be determined gravimetrically. In case of water samples this can be achieved by collecting the water sample in a separatory funnel that contains already the required volume of solvent.

DATA AND REPORTING

• 47.According to the present Test Method, POW is determined by performing three slow-stirring experiments (three experimental units) with the compound under investigation employing identical conditions. The regression used to demonstrate attainment of equilibrium should be based on the results of at least four determinations of CO/CW at consecutive time points. This allows for calculating variance as a measure of the uncertainty of the average value obtained by each experimental unit.

• 48.The POW can be characterized by the variance in the data obtained for each experimental unit. This information is employed to calculate the POW as the weighted average of the results of the individual experimental units. To do so, the inverse of the variance of the results of the experimental units is employed as weight. As a result, data with a large variation (expressed as the variance) and thus with lower reliability have less influence on the result than data with a low variance.

• 49.Analogously, the weighted standard deviation is calculated. It characterizes the repeatability of the POW measurement. A low value of the weighted standard deviation indicates that the POW determination was very repeatable within one laboratory. The formal statistical treatment of the data is outlined below.

Treatment of the results

Demonstration of attainment of equilibrium

• 50.The logarithm of the ratio of the concentration of the test substance in 1-octanol and water (log (CO/Cw)) is calculated for each sampling time. Achievement of chemical equilibrium is demonstrated by plotting this ratio against time. A plateau in this plot that is based on at least four consecutive time points indicates that equilibrium has been attained, and that the compound is truly dissolved in 1-octanol. If not, the test needs to be continued until four successive time points yield a slope that is not significantly different from 0 at a p-level of 0,05, indicating that log Co/Cw is independent of time.

Log POW-calculation

• 51.The value of log POW of the experimental unit is calculated as the weighted average value of log Co/Cw for the part of the curve of log Co/Cw vs. time, for which equilibrium has been demonstrated. The weighted average is calculated by weighting the data with the inverse of the variance so that the influence of the data on the final result is inversely proportional to the uncertainty in the data.

Average log POW

• 52.The average value of log POW of different experimental units is calculated as the average of the results of the individual experimental units weighted with their respective variances.

• The calculation is performed as follows:

Image omitted; see official source

• where:

log POW,i	=	the log POW value of the individual experimental unit i;
---	---	---
log POW,Av	=	the weighted average value of the individual log POW determinations;
---	---	---
wi	=	the statistical weight assigned to the log POW value of the experimental unit i.

• The reciprocal of the variance of log POW,i is employed as wi (

Image omitted; see official source )

• 53.The error of the average of log POW is estimated as the repeatability of log Co/Cw determined during the equilibrium phase in the individual experimental units. It is expressed as the weighted standard deviation of log POW,Av (σlog Pow,Av) which in turn is a measure of the error associated with log POW,Av. The weighted standard deviation can be computed from the weighted variance (varlog Pow,Av) as follows:

Image omitted; see official source

Image omitted; see official source

• The symbol n stands for the number of experimental units.

Test Report

• 54.The test report should include the following information:

Test substance:

• — common name, chemical name, CAS number, structural formula (indicating position of label when radiolabelled substance is used) and relevant physical-chemical properties (see paragraph 17)

• — purity (impurities) of test substance

• — label purity of labelled chemicals and molar activity (where appropriate)

• — the preliminary estimate of log Pow, as well as the method used to derive the value.

Test conditions:

• — dates of the performance of the studies

• — temperature during the experiment

• — volumes of 1-octanol and water at the beginning of the test

• — volumes of withdrawn 1-octanol and water samples

• — volumes of 1-octanol and water remaining in the test vessels

• — description of the test vessels and stirring conditions (geometry of the stirring bar and of the test vessel, vortex height in mm, and when available: stirring rate) used

• — analytical methods used to determine the test substance and the method limit of quantification

• — sampling times

• — the aqueous phase pH and the buffers used, when pH is adjusted for ionizable molecules

• — number of replicates.

Results:

• — repeatability and sensitivity of the analytical methods used

• — determined concentrations of the test substance in 1-octanol and water as a function of time

• — demonstration of mass balance

• — temperature and standard deviation or the range of temperature during the experiment

• — the regression of concentration ratio against time

• — the average value log Pow,Av and its standard error

• — discussion and interpretation of the results

• — examples of raw data figures of representative analysis (all raw data have to be stored in accordance with GLP standards), including recoveries of surrogates, and the number of levels used in the calibration (along with the criteria for the correlation coefficient of the calibration curve), and results of quality assurance/quality control (QA/QC)

• — when available: validation report of the assay procedure (to be indicated among references).

LITERATURE:

• (1) De Bruijn JHM, Busser F, Seinen W, Hermens J. (1989). Determination of octanol/water partition coefficients with the ‘slow-stirring’ method. Environ. Toxicol. Chem. 8: 499-512.

• (2) Chapter A.8 of this Annex, Partition Coefficient.

• (3) Chapter A.8 of this Annex, Partition Coefficient.