ISO 20846 Petroleum products - Determination of sulfur content of automotive fuels - Ultraviolet fuorescence method
6 Sampling
6.1 Unless otherwise specifed, obtain the laboratory sample by the procedures described in ISO 3170 for manual sampling, or ISO 3171 for automatic pipeline sampling. To preserve volatile components found in some samples, keep samples at as low a temperature as is practicable and do not uncover samples any longer than necessary.

Samples should be analysed as soon as possible after removal from bulk supplies, to prevent loss of sulfur or contamination due to exposure or contact with the sample container.

CAUTION - Samples that are collected at temperatures below room temperature can undergo expansion and rupture the container. For such samples, do not fill the container to the top; leave suffcient air space above the sample to allow room for expansion.

6.2 If the sample is not used immediately, thoroughly mix it in its container prior to taking a test portion.

7 Apparatus preparation
7.1 Assemble the apparatus and check for leaks according to the manufacturer's instructions.

7.2 Adjust the inlet pressure and fow rate of each gas according to the manufacturer's instructions.

7.3 Refer to the manufacturer's instructions to set the temperature of the furnace (5.1) high enough to ensure all sulfur pyrolysis and oxidation, typically 1 100 °C in the case of a one temperature zone furnace, or 750 °C for pyrolysis, and 1000 °C to 1100 °C for oxidation in the case of a two temperature zone furnace.

7.4 Adjust the instrument sensitivity and baseline stability, and perform instrument blanking procedures following the manufacturer's guidelines. Ensure that the UV light is stable before measurement.

NOTE For the UV light, a warm-up time of at least 30 min is usually required.

8 Apparatus calibration and verifcation
8.1 Multi-point calibration
8.1.1 Select one of the suggested curves outlined in Table 1. Prepare a series of calibration standards (4.6) by carrying out dilutions of the stock solution (4.5) with the selected solvent (4.3) to cover the range of operation. The number of calibration standards used in construction of the calibration curve can vary, but it should not be less than four.

8.1.2 Flush the microlitre syringe (5.6) several times with the solution to be analysed and ensure that the fnal liquid column in the syringe contains no bubbles.

8.1.3 Using one of the techniques described 8.1.3.1 or 8.1.3.2, quantitatively analyse an appropriate sample size as specifed by the manufacturer, prior to injection into the combustion tube for analysis.

NOTE Injection of a constant or similar sample size for all materials analysed in a selected operating range promotes consistent combustion conditions.

8.1.3.1 For volumetric measurement, fll the syringe to the selected level and retract the plunger so that air is aspirated and the lower liquid meniscus falls on the 10 % scale mark. Record the volume of liquid in the syringe. After injection, again retract the plunger so that the lower liquid meniscus falls on the 10 % scale mark and record the volume of liquid in the syringe. The difference between the two volume readings is the volume of sample injected.

NOTE An automatic sampling and injection device can be used in place of the described manual injection procedure.

8.1.3.2 For mass measurement, weigh the syringe complete with a flled needle before injection, and the syringe and needle after injection, to determine the mass of the test portion injected.

NOTE Mass measurement can be more accurate than the volume measurement for less volatile samples, if a balance with an accuracy of at least +/- 0.01 mg is used.

8.1.4 Once the appropriate sample size has been measured into the microlitre syringe, promptly and quantitatively deliver the sample into the apparatus.

For direct injection, carefully insert the syringe into the inlet of the combustion tube (5.2) and the syringe drive. Allow time for sample residues to be burned from the needle (needle blank). Once a stable baseline has been re-established, promptly start the analysis. Remove the syringe once the apparatus has returned to a stable baseline.

8.1.5 Construct each calibration curve by one of the techniques described in 8.1.5.1 or 8.1.5.2.
8.1.5.1 For manual construction, analyse the calibration standards and the blank three times using the procedure described in 8.1.2 to 8.1.4. Subtract the average blank response from each analysis of the standard before determining the average integrated response. Construct a curve plotting average integrated detector response (y-axis) versus quantity, Q, expressed in nanograms, of sulfur injected (x-axis). This curve should be linear with a correlation factor of at least 0.995.

Q is calculated using Equation (1) or (2):
Q = mc x wSc
Q =Vc x ρS
where
wSc is the sulfur content of the calibration standard, expressed in milligrams per kilogram;
Vc is the volume of the calibration standard injected, expressed in microlitres;
ρS is the sulfur concentration of the calibration standard, expressed in milligrams per litre;
mc is the mass of the calibration standard injected, expressed in milligrams, either measured directly, or calculated from the measured volume injected and the density using Equation (3):
mc = Vc x Dc
where
Dc is the density of the calibration standard at measurement temperature, expressed in grams per millilitre.

8.1.5.2 If the apparatus features an internal calibration routine, analyse the calibration standards and the blank three times using the procedure described in 8.1.2 to 8.1.4. If blank correction is required and is not available (see 4.3), correct the analyser responses using the average response for each standard versus quantity, Q (in nanograms), of sulfur as obtained in 8.1.5.1. This curve should be linear with a correlation factor of at least 0.995.

NOTE For curve III (Table 1), the blank correction can be omitted.

8.1.6 The analyser calibration can be performed using a different calibration curve from those listed in Table 1. It is good practice to construct a calibration curve in such a way that the sulfur content of the sample under analysis is in the middle of the calibration curve.

8.2 One-point calibration
8.2.1 Prepare a calibration standard (4.6) with a sulfur content close to that of the sample under analysis (+/- 50 % maximum), by carrying out dilutions of the stock solution (4.5). A calibration standard with a sulfur concentration greater than 50 % from that of the sample may be used if the linearity of the apparatus has previously been checked.

If necessary, the sulfur content of the calibration standard should be corrected by the sulfur content of the selected solvent (4.3).

NOTE In order to measure the sulfur content of the selected solvent, the method of standard additions of the solvent gives satisfactory results.

8.2.2 Using the procedure described in 8.1.2 to 8.1.4, analyse a minimum of three times the calibration standard using an appropriate sample size as specifed by the manufacturer.

8.2.3 Calculate a calibration factor, K, expressed in number of counts per nanogram of sulfur, using Equation (4) or (5):
K = Ac/(mc x wSc)
K = Ac/(Vc x ρS)
where
Ac is the integrated detector response for calibration standard, in number of counts;
wSc is the sulfur content of the calibration standard, expressed in milligrams per kilogram;
Vc is the volume of the calibration standard injected, expressed in microlitres;
ρS is the sulfur concentration of the calibration standard, expressed in milligrams per litre;
mc is the mass of the calibration standard injected, expressed in milligrams, either measured directly or calculated from the measured volume injected and the density using Equation (6):
mc = Vc x Dc
where
Dc is the density of the calibration standard at measurement temperature, expressed in grams per millilitre.

Calculate the average of the calibration factor and check that the standard deviation is within the accepted tolerance.

When the concentration of calibration standards is known in milligrams per kilogram, and volume injections are performed, a correction for the difference of density between standards and samples is necessary.

8.3 Verifcation
8.3.1 Check the calibration validity with a quality control sample of known sulfur content, at the beginning of a series of analyses and at least every twenty analyses during a long series of analyses.

8.3.2 Compare the results to the known values and associated uncertainties. The out of limit results shall be
investigated for root cause(s).

NOTE 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 method, or 0.7 times the reproducibility, are reasonable starting points.