ASTM D7757 Standard Test Method for Silicon in Gasoline and Related Products by Monochromatic Wavelength Dispersive X-ray Fluorescence Spectrometry
9. Sampling and Sample Handling
9.1 Sample fuel according to the procedures in Practices D4057 or D4177.

9.2 Use the utmost care in sampling and handling gasoline to prevent evaporation of light ends which could change the concentration of silicon in the sample. Store gasoline in a leak tight container at 0 °C to 4 °C until ready for analysis. If possible, maintain at this temperature throughout any transfer and handling processes. Allow samples maintained at 0 °C to 4 °C to come to room temperature before testing, and expose these materials to ambient conditions only as long as necessary to obtain a sample for analysis. Analyze test specimens as soon as possible after sub-sampling from bulk container. Do not allow bulk container to remain uncovered any longer than is needed to obtain desired sub-samples.

9.3 For specimen preparation, see 10.2.
9.3.1 Because impurities and thickness variations can occur in commercially available transparent films and vary from lot to lot, use calibration-check samples (see 7.2) to verify calibration integrity after starting each new batch of film or if the type and thickness of the window film is changed.

9.4 When reusable sample cells are used, thoroughly clean and dry cells before each use.
9.4.1 Disposable sample cells shall not be reused.

10. Preparation of Apparatus and Specimens for Analysis
10.1 Analyzer Preparation - Ensure that the MWDXRF analyzer has been installed and put into operation according to manufacturer's instructions. Allow sufficient time for instrument electronics to stabilize. Perform any instrument checkout procedures required. When possible, the instrument should be run continuously to maintain optimum stability.
10.1.1 Use the count time (T) recommended by the instrument manufacturer for the lowest silicon concentration expected. The typical time for each measurement is five to ten minutes.

10.1.2 Alternatively, determine T expected for a desired count precision by following the procedure in Appendix X1.

10.2 Specimen Preparation - Prepare a specimen of a test sample or a calibration standard as follows:
10.2.1 Carefully transfer a sufficient portion of the liquid to fill a sample cell above a minimum depth beyond which additional liquid does not affect the count rate. In a XRF sample cup, filling the sample cell to two thirds of the cell's depth is generally adequate.

10.2.2 Fit an unused piece of X-ray-transparent film over the sample-cell opening and attach securely. Use the same batch offilm for the analysis oftest samples and the calibration standards used for constructing the calibration curve. Avoid touching the inside of the sample cell, any portion of the film exposed to the liquid or the X-ray beam, and also avoid touching the instrument window. (It is highly recommended that clean, disposable rubber or plastic gloves be used when preparing test specimens.) Oil from fingerprints can generate errors in the analysis of silicon. Ensure the film is clean, taut, wrinkle-free, and the test specimen is not leaking to ensure reliable results.

10.2.3 Provide a small vent to prevent bowing of the window film caused by the accumulating vapor. Many commercially available sample cells provide a means to vent the space above the liquid.

10.2.4 Perform the analysis of the specimen promptly after preparing the specimen. Do not let the specimen remain in the sample cell any longer than necessary before collecting the data.

11. Calibration
11.1 Obtain or prepare a set of calibration standards by careful mass dilution of octamethylcyclotetrasiloxane (D4) with a suitable base material (BM) (see Section 5). The concentrations of the unknown samples shall lie within the calibration range that is used. All standards used in the analysis shall be from a reliable and consistent source, which may include commercially available standards. Approximate recommended nominal silicon concentration standards are listed as follows for the range of 3 mg/kg to 100 mg/kg: 0.0 mg/kg (base material), 10 mg/kg, 25 mg/kg, 100 mg/kg, and 250 mg/kg.
11.1.1 Take into account any silicon in the base materials when calculating the silicon content (mg/kg) in each of the calibration standards as shown in Eq 1:
Si = [(D4*SiD4) + (BM*SiBM)]/(D4 + BM)
where:
Si = mass fraction of silicon in the prepared standards, mg/kg,
D4 = actual mass of octamethylcyclotetrasiloxane, g,
SiD4 = mass fraction of silicon in D4, mg/kg, typically 37.838%,
BM = actual mass of base material, g, and
SiBM = mass fraction of silicon in the base material, mg/kg.

11.2 Following instrument manufacturer's instructions and the instructions in 12.2, measure the silicon fluorescence intensity (total silicon count rate) for each of the calibration standards. Convert total counts (N) to count rate (RS) in counts per second by dividing N by the count time (T) using units of seconds (see 10.1.1, 10.1.2, and Eq 2).
RS = N/T
RS = measured total count rate of the silicon fluorescence from 11.2, counts per second,
N = total counts collected at 0.713 nm, and
T = seconds required to collect N counts.

11.3 Construct a linear calibration model by either:
11.3.1 Using the software supplied by the instrument manufacturer, or

11.3.2 Perform a linear regression of the calibration measurements. The following linear equation (Eq 3) describes the regression:
RS = Y + E x Si
RS = measured total count rate of the silicon fluorescence from 11.2, counts per second,
Y = y-intercept of the calibration curve, counts per second,
E = slope of the calibration curve, counts kg s(-1)mg(-1), and
Si = silicon concentration, mg/kg.

11.4 When using drift correction, measure the total counts of silicon fluorescence from the drift-monitor sample during the calibration procedure. Determine R S by dividing the total counts by T. The factor, RS, determined on the drift-monitor sample at the time of calibration, is factor A in Eq 4.

11.5 Immediately after analyzing the calibration standards, determine the silicon concentration of one or more calibration-check samples (see 7.2). The determined value shall be in the range defined by the certified concentration plus or minus the repeatability of this test method. If this criterion is not met, the calibration process and calibration standards are suspect, corrective measures shall be taken, and the calibration repeated. The degree of matrix mismatch between calibration check samples and standards should be considered when evaluating a calibration.