ISO 20884 Petroleum products - Determination of sulfur content of automotive fuels - Wavelength-dispersive X-ray fluorescence spectrometry
6 Sampling
Unless otherwise specified, samples shall be taken by the procedures described in ISO 3170 or ISO 3171.

7 Calibration solutions
7.1 Blank solution
Use white oil (4.2) as a blank solution.

If FAME is under analysis, use a blank solution in accordance with 4.3 to minimize potential matrix effects.

7.2 Stock solution
Weigh a quantity of the required calibrating substance (4.1) to the nearest 0.1 mg, and dilute with blank solution (4.2 or 4.3, as appropriate) at room temperature so that a stock solution with an accurately known sulfur content (to the nearest 1 mg/kg) of approximately 1000 mg/kg is obtained.

Care should be taken due to volatility of the calibrating substance (see Annex A).

7.3 Calibration solutions
Weigh an appropriate quantity of the stock solution (7.2) to the nearest 0.1 mg into bottles and dilute with blank solution (4.2 or 4.3, as appropriate) so that standard solutions with sulfur contents in accordance with Tables 2 and 3 are obtained by stirring at room temperature. The sulfur contents in the calibration solutions are indicated in milligrams per kilogram, rounded to the nearest 0.1 mg/kg.

7.4 Storage and stability of the calibration solutions
Calibration solutions prepared in accordance with Table 2 have a limited stability and shall be used on the day they are prepared.

Calibration solutions prepared in accordance with Table 3 have a stability of no more than one week if stored in a cool location (refrigerator).

8 Settings
8.1 Measuring parameters
For optimum measuring parameters, refer to Table 1.

8.2 Optimization
The spectrometer shall be optimized in accordance with the manufacturer's specifications in order to achieve the optimum signal-to-noise ratio. The calibration solution with 50 mg/kg sulfur is recommended for optimization.

The counting time shall be set so that the total net count from a 50 mg/kg solution is ≥ 40000. This counting time shall be used for both calibration (see Clause 9) and measurement (see Clause 10).

8.3 Performance check of the spectrometer
Prior to a measurement series (calibration and/or measurement), but in any case at least once daily, the manufacturer's specifications shall be used to check that the spectrometer is operating correctly, so that an optimum performance and a consistent standard of highest possible quality is ensured.

Performance checks should be carried out on a regular basis, since such checks give valuable information about the status and stability of the spectrometer.

9 Calibration
9.1 General
Ensure that the spectrometer is in an optimized condition after having executed all provisions given in Clause 8 and operating optimally before calibration (see 8.3).

For the range from 5 mg/kg to 60 mg/kg sulfur, and for the range 60 mg/kg to 500 mg/kg sulfur, separate calibrations shall be conducted.

9.2 Calibration solutions
Transfer the calibration solutions (7.3) into suitable sample cups so that the minimum volume is achieved (see paragraph below). Depending on the concentration range, all calibration solutions prepared in accordance with Tables 2 or 3 shall be measured consecutively in ascending order. The count rate, Is, of the SK-L2,3 X-ray fluorescence radiation at 0.5373 nm, and the count rate, IB, of the background radiation at 0.545 nm, shall be measured consecutively. For instruments utilizing monochromatic excitation, measurement of the background radiation is not required (see 5.1).

Sample quantities that are too small give a low result or relatively more evaporation/concentration effects when measuring volatile samples, while too much sample causes more sagging of the cell window, especially when light aromatic samples are measured. It is necessary to determine the minimum volume by proceeding as follows: start with 1 ml, measure the net count rate, add liquid, measure again the count rate. Repeat until the count rate remains constant for greater volumes of liquid. Plot a curve of net count rate vs volume and choose a volume greater than the point at which the curve becomes flat.

9.3 Calibration curves
The net count rate, R0, is calculated according to Equation (1) or (2). To determine the calibration, the net count rate, R0, is plotted as a function of the corresponding sulfur content. Use the spectrometer software or a spreadsheet program to perform a regression of the data to fit to Equation (3).

The regression calculation may be carried out either separately or using the spectrometer calculator.
R0 = IS − IB when background is measured, or
R0 = IS for monochromatic excitation, when no background is measured
R0(x) = a + bx + cx2
where
R0 is the net count rate for the relevant determination;
IS is the count rate of the SK-L2,3 X-ray fluorescence radiation at 0.5373 nm;
IB is the count rate of the background radiation at 0.545 nm;
x is the sulfur content of the relevant calibration solution, expressed in milligrams per kilogram (mg/kg);
a, b, c are the parameters from the regression calculation;
R0(x) is the net count rate computed from the regression calculation for the content, x.

NOTE Parameter c may be zero where the calibration curve has been shown to be linear.

9.4 Checking
Check at least two points on each calibration curve regularly, but not less frequently than every six months. It is highly recommended that quality control samples be used, with known sulfur contents for these checks. Check immediately when using a new batch of films. If the check result differs from the curve, execute a new calibration. Control limits are established from the laboratory statistical control charts, but initial values should be set before experience is established. Limits of the repeatability of this International Standard, or 0.7 times the reproducibility, are reasonable starting points. In case of suspected changes in the equipment status (e.g. drift), recalibration is required.