ASTM D6376 Standard Test Method for Determination of Trace Metals in Petroleum Coke by Wavelength Dispersive X-ray Fluorescence Spectroscopy
9. Sampling and Preparation of Test Specimens
9.1 Basic Assumptions - All petroleum coke samples examined using this test method shall be analytical samples representative of a green or calcined coke production lot or shipment. Practices D4057, D346, D6969, and D6970 shall be followed. The analytical sample shall pass a 75 µm (U.S. No. 200 mesh) sieve and shall be of sufficient quantity to provide sample for all testing and analyses planned for the particular lot of petroleum coke. To provide for possible replicate determinations of the sulfur and metallic contents by this test method and for a retain sample, an analytical sample shall be 50 g or more for this test alone.

NOTE 1 - Results are particle dependent, and erroneous data may be collected if analytical samples contain particles varying significantly in size from those in reference samples. Measurements may be necessary to ensure the equivalence of analytical samples and reference samples. The most difficult problems occur when there are coarse petroleum coke samples and finely ground reference samples. In this case, it is best to grind the analytical samples, achieving a size distribution similar to the reference samples.

9.2 Types of Samples:
9.2.1 Reference Samples - These samples shall be similar to the analytical samples in composition and physical properties. A wide range of concentrations is required for inter-element effect detection when regression software is used to analyze reference data. Elemental concentrations shall bracket the values expected in the analytical samples.

9.2.2 Analytical Samples - Samples whose sulfur and metallic contents are to be determined.

9.3 Typical Preparation ofStandards and Test Specimens:
9.3.1 Dry a sufficient portion of the analytical sample to constant mass in an oven at 110 °C +/- 10 °C to obtain at least 5 g.

9.3.2 Weigh 5 g +/- 0.01 g of the dried analytical sample.

9.3.3 Add 1 g +/- 0.01 g of binder and mill for a minimum of 20 s. Binder/coke ratio shall remain constant for samples and standards.
9.3.3.1 Suitable grinding times to produce particles passing a 75 µm (U.S. No. 200 mesh) sieve depends on the type of grinder used and coke variations. Further grinding time impacts element intensities. It is important that grinding time be thoroughly investigated and optimized.

NOTE 2 - It is essential that the same sample preparation procedure (including sample mass, binder mass and ratio, grinding, and so forth) be followed precisely for all analytical and reference samples. Even a small change in procedure requires making all new reference samples match the changed procedure. All reference samples and analytical samples used with them shall be prepared in exactly the same manner. All weighings are to be made to the nearest 0.01 g.

9.3.4 Place the pellet cup in the die mounted on the hydraulic press. Transfer sufficient ground mix to this cup to produce a pile of maximum height above the cup lip. Hand pack with a flat spatula until the ground mix is level with the cup lip.

9.3.5 Apply adequate pressure to achieve a stable pellet. Allow adequate time to reach a pressure of approximately 276 MPa (40 000 psi) and hold for a minimum of 5 s.

9.3.6 Release the pressures slowly, and remove the pellet from the die. (Warning - Release the pressure slowly to avoid possible damage to the pressure gauge.)

9.3.7 Inspect the pellet surface to ensure that it is smooth and free of cracks. The cylindrical pellets should be 3 mm to 7 mm thick.

9.3.8 When the pellet is too thin, repeat 9.3.2-9.3.7 using about 20 % more ground mix.

9.3.9 Clean the outside surface of the pellet cup, using a clean cloth or tissue. If the pellet is to be stored, it is important to properly identify it and store it in a dry place.

9.3.10 Unused test specimens prepared and stored as above are generally stable for years. Reference sample pellets can be used daily for establishing instrument conditions. If changes greater than 3 % (10 % for silicon) are observed for the calculated concentrations for any element in a used pellet, the reference sample pellets shall be prepared again. If deviations persist, recalibrate the instrument.

10. Preparation of Apparatus
10.1 Follow manufacturer's instructions for the initial assembly, conditioning, and preparation of the XRF unit.

10.2 Follow the manufacturer's control setting and operation instructions.

10.3 Peak and Background Intensity Measurements - A decision on counting time is made after analyzing all required reference samples and after the sensitivity is known. A suggested strategy for counting time based on reference samples is as follows:
10.3.1 Counting time is derived from:
Relative error, % = 100/(Tt)½ x 1/((Rp)½ - (Rb)½)
where:
Tt = the total counting time for the peak and the background. T t can be calculated as all other terms, either known or measured,
Rp = the count rate for the peak, and
Rb = the count rate for the background.

10.3.2 The partitioning of counting time based on the peak and background is estimated from:
Tp/Tb = (Rp/Rb)½
where:
Tp = the peak count time,
Tb = the background count time, and
Rp and R b = values used to reflect intensity data from a mid to low concentration standard reference material.

11. Reference Samples and Calibration
11.1 Samples Used for Calibration:
11.1.1 Reference samples for calibration can be purchased or prepared in-house.

NOTE 3 - It is recommended that reference samples used for calibration be verified by Test Method D1552 for sulfur, by Test Method D5600 or D5056 for metals, or by using other techniques considered standards in the industry, such as spectrophotometry.

11.1.2 Reference sample concentrations shall bracket the concentrations of the analytical samples. The minimum number of reference samples is as follows:
No. of stds = 3 (2 + N)
where:
N = the number of corrections calculated by the regression analysis program.
11.1.2.1 Using too few reference samples may result in lines that are over defined, that is, sulfur and vanadium, which can be highly correlated.

11.2 Calibration:
11.2.1 Collect data using the procedure in Section 12.

11.2.2 Sulfur correction factors, the slope, and the intercept of the calibration line are obtained by regression analysis using the spectrometer software or a model similar to the following:
Ci = (Di + EiRi) (1 + αisCs)
where:
Ci = concentration of the analyte element i,
Di = intercept of the calibration curve for element i,
Ei = slope of the calibration curve for element i,
Ri = measured net intensity for element i,
αis = interelement correction factor for the effect of sulfur on analyte element i, and
Cs = concentration of sulfur.

11.2.3 A slope, an intercept, and a sulfur correction factor shall be calculated for each metallic element.

11.2.4 Sulfur correction factors (αis) can also be calculated from special software.

11.2.5 A stable pellet (monitor sample) shall be used periodically to monitor instrument drift. A change in the drift factor by more than +/- 10 % indicates a major change or problem has occurred in the equipment. Checks and possibly recalibration may be necessary. The drift factor d is calculated for each element by the ratio:
d = R1/Rn
where:
R1 = intensity of monitor sample during the calibration procedure, and
Rn = intensity of monitor sample when unknown samples are measured.