ASTM D7578 Standard Guide for Calibration Requirements for Elemental Analysis of Petroleum Products and Lubricants
7. Calibration in Empirical Test Methods
7.1 In this guide, empirical methods are defined as the methods for which no special calibration of instrumentation is necessary or if needed is minimal. Examples of such analysis include Test Methods D129 (Sulfur by Bomb Method), D482 (Ash), D808 (Chlorine by Bomb Method), D874 (Sulfated Ash), D1018 (Hydrogen), D1091 (Phosphorus), D1266 (Sulfur by Lamp Method), D2784, A (Sulfur), D3228 (Kjeldahl Nitrogen), D4047 (Phosphorus), and D5384 (Chlorine). In most of these cases either gravimetric or titrimetric finish are required. Thus, only calibrations involved will be for analytical balance, temperature probe for ashing steps, and standardization of titrants, where required.

8. Calibration in Photometric Test Methods
8.1 A number of test methods use flame emission or colorimetric measurements for quantitation of analytes, usually after reacting the matrix with a chromogenic reagent. Such instruments need to be calibrated using appropriate standards developed by reacting the pure metal analyte with the chromogenic reagent. The calibration curve of analyte concentration versus the photo-signal (absorbance or transmittance) should follow Beer-Lambert Law. The sample analysis should be carried out only in the linear range of the plot. If necessary, solutions with higher concentrations of metals should be diluted to bring them into the linear range of calibration.

8.2 Generally, between three and eight standards are used for developing the calibration curve. The absorbance or transmittance of standard solutions is corrected by subtracting the signal from the blank solution. If a chemical procedure is used to develop the color, the same reagents and steps need also to be included in developing the blank color.

8.3 The test methods using such photometric methods appear in Table 4.

9. Calibration in Electrometric Test Methods
9.1 A number of elemental analysis test methods are based on final measurements using electrometric techniques such as use of microcoulometry, thermal conductivity, chemiluminescence, UV-fluorescence, etc. Most of these test methods are used for the determination of non-metals such as halogens, oxygen, sulfur, carbon-hydrogen-nitrogen, as well as anions such as sulfate and chloride.

9.2 Three to six calibration standards are usually used to construct a response curve for the analysis. Specific requirements are given in each ASTM Standard Test Methods listed below. Mostly these calibrants are pure compounds. Some examples of such compounds are given here, although other appropriate compounds may be used for calibration if they are shown to produce equivalent results as the following:
Nitrogen: Carbazol, Pyridine, Acridine
Sulfur: Di-Butyl Sulfide, Dibenzothiophene, Thionaphthene, Dimethyl Sulfide
Chlorine: Chlorobenzene, Sodium Chloride
Oxygen: Anhydrous Methanol, NIST SRM 1837
Sulfate: Anhydrous Sodium Sulfate
C-H-N: Acetanilide, Atropine, etc.

9.3 A calibration curve with at least three standards is recommended, covering the range of analyte concentrations expected in the samples being analyzed. Calibration curves with one or two standards can be used if it can be shown that the calibration curve is linear over the concentration range being measured and the standards meet the performance criteria given in the test method. Larger series of standards can also be used if it can be shown that the calibration curve is linear over the concentration range being measured and the standards meet the performance criteria of the test method.

9.4 Since these test methods often measure the elements at very low level of concentration, the pure organic compounds used shall be dissolved in appropriate organic solvents and diluted to bring the concentration down in the working linear range (usually in mg/kg or μg/mg).Clean glassware shall be used for dilution and storage purposes. Preparation of working and calibration standards on a regular basis is recommended, depending upon the frequency of usage and age. Stock solutions and particularly the working standards may have a limited useful life. Stock, working, and calibration standards from commercial sources can be used if checked for accuracy and can meet the performance criteria given in the test methods.

9.5 The tests methods using electrometric finish appear in Table 5.

10. Calibration in Atomic Absorption Spectrometry
10.1 A number of metals are determined by the popular technique AAS. In all cases the instrument needs to be calibrated before each set of analysis. Usually solutions of organometallic standards dissolved in solvents such as xylene, toluene, MIBK, kerosene, etc. are utilized. Such standards are widely available from commercial sources or NIST. The concentration of metals in the standards for actual calibration is usually kept between 1 to 10 mg/kg, prepared by dilution from stock solutions. Although the purchased stock solutions of high metal concentrations can be preserved for long periods of time, the diluted working standards should not be kept for more than a day, since it is possible that there may be loss of analyte by adsorption on the container walls.

10.2 Similar to the photometric methods, a plot or correlation based on metal concentration versus absorbance is prepared and the metal concentration in the samples is calculated based on this factor. The analysis should be done only within the range of linear curve. If expected metal concentration is higher than the linear range, the sample should be diluted with appropriate solvent otherwise results biased low will result.

10.3 The calibration curve may be manually plotted using x-axis for concentration of metal in the working standards versus corrected absorbance on the y-axis. Many AAS instruments have the capability of automatically constructing the calibration curve internally by way of the instrument software and displayed by way of the instrument computer terminal, making actual manual plotting unnecessary. A curve with the best possible fit of the data within the available means should be used

10.4 Most modern AAS instruments can store up to three or four calibration standards in memory. In such cases, follow the manufacturer's instructions, ensuring that the unknown sample's absorbance is in the linear part of the calibration range used.

10.5 Generally stock solutions of standards contain 100 to 1000 mg/kg metal concentration. For working standards these are diluted with appropriate organic solvents to a level of about 10 or lower mg/kg level.

10.6 Generally three to six standards are used for establishing the calibration curve of metal concentration versus absorption signal. Often though it has been found that one standard in the linear range is sufficient. A blank needs to be subtracted from the standard solutions' absorbance for compensating any contribution from the metals contaminating the standard solutions. If any chemical step is involved in final standard preparation, it should also be included in the blank solution preparation.

10.7 Calibration shall be carried out prior to each group of samples to be analyzed and after any change in instrumental conditions, as variation occurs in the instrument behavior. Readings also may vary over short periods of time from such causes as buildup of deposits on the burner slot or in the nebulizer. Thus, a single standard should be aspirated from time to time during a series of samples to check whether the calibration has changed. A check after every 5th sample is recommended. The visual appearance of the flame also serves as a useful check to detect changes of condition.

10.8 Although the AAS methods are depicted as single element analysis methods, they can be used for the determination of multiple elements in a single sample. In such cases the sequence of operations in analyzing several samples should also be considered. Aspiration of a sample to determine the absorbance is very quick. Changing wavelength setting and lamps takes longer. Thus, it is most economical to make measurements at a single wavelength on a series of samples and standards before changing conditions. In such cases use of multi-element calibration standards and multi-element hollow cathode lamps would be useful.

10.9 Some of the test methods available for petroleum products and lubricants analysis using AAS appear in Table 6.

10.10 Although not strictly the AAS or ICP-AES type methods, D6595 and D6728 use rotating disc electrode atomic emission spectrometry methods for the determination of wear metals and contaminants in used lubricating oils, used hydraulic fluids, and gas turbine and diesel engine fuels. The calibration operation range of such instruments for each element is established through the analysis of organometallic standards at known concentrations in the instrument manufacturing factory. A calibration curve for each element is established and correction factors are set to produce a linear response. Analyses of the test specimen shall be performed within the linear range of the response. A minimum of a two point routine standardization should be performed if the instrument fails the validation check or at the start of each working shift. A minimum of three analyses should be made using the blank and working standard.

11. Calibration in ICP-AES Test Methods
11.1 Given the capability of ICP-AES for simultaneous multi-element capability, most laboratories use multi-element standards for instrument calibration. Although such standards can be prepared in-house from pure organometallic standards, it is much more practical and convenient to use pre-blended multi-element organic standards available from many commercial sources. Most are available dissolved in base oil or other organic solvents at levels from 20 to 1000 mg/kg individual element levels. Custom blended standards can also be obtained from these sources.

11.2 More than one multi-element standard may be necessary to cover all elements of interest. It is imperative that concentrations are selected such that the emission intensities measured with the working standards can be measured precisely (that is, the emission intensities are significantly greater than background), and these standards represent the linear region of the calibration curve. Frequently, the instrument manufacturer publishes guidelines for determining linear range.

11.3 Some commercially available organometallic standards are prepared from metal sulfonates, and therefore contain sulfur. Thus, for sulfur determination a separate sulfur calibration standard can be required. Metal sulfonates can be used as a sulfur standard if the sulfur content is known or determined by an appropriate test method such as D1552. Petroleum additives can also be used as organometallic standards if their use does not adversely affect the precision nor introduces significant bias.

11.4 Some of the ICP-AES methods (for example, D5184, D5600, and D7303) use aqueous solutions after decomposing the organic matrices converting them into dilute acidic aqueous solutions. In such cases aqueous metal standards shall be used. These are widely available from commercial sources usually in 100 to 1000 mg/kg concentrations. They also can be prepared in-house by dissolving inorganic metal salts in dilute acids.

11.5 The linear range of all ICP-AES curves shall be determined for the instrument being used. This is accomplished by running intermediate standards between the blank and the working standards and by running standards containing higher concentrations than the working standards. Analyses of test specimen solutions shall be performed within the linear range of the calibration curve. At the beginning of the analysis of each set of test specimen solutions, a two-point calibration using the blank and the working standard is performed. A check standard is used to determine if each element is in calibration. When the results obtained with a check standard are within 5 % relative of the expected concentration for all elements, the analysis may be continued. Otherwise, any necessary adjustments to the instrument need to be made and the calibration repeated.

11.6 The calibration curves can be constructed differently, depending on the implementation of internal standard compensation.
11.6.1 When analyte intensities are ratioed to internal standard intensities, the calibration curve, in effect, is a plot of intensity ratio for analyte versus the analyte concentration.

11.6.2 When the internal standard compensation is handled by multiplying all results for a test specimen by the ratio of the actual internal standard concentration to the determined internal standard concentration, the calibration curve is, in effect, a plot of (intensity for analyte - intensity of the blank for the analyte) versus analyte concentration.

11.7 Although normally the calibration curves would be linear, sometimes inclusion of a second order term can give a better fit, although in such cases it would require at least five standards for calibration.

11.8 Detailed discussion of calibration and all other pertinent protocols for use of ICP-AES for analyzing petroleum products and lubricants is given in Practice D7260.

11.9 Some of the test methods available for analysis of petroleum products and lubricants using ICP-AES appear in Table 7.

12. Calibration in XRF Test Methods
12.1 Similar to ICP-AES, XRF has the capability of multi-element analysis, even with a better precision. A large number of XRF test methods have been written for the determination of low levels of sulfur in fuels because of its importance in environmental and regulatory affairs.
12.1.1 Detailed discussion about the sample handling, calibration and validation using XRF methods can be found in Practice D7343.

12.2 Some of the test methods utilizing XRF technique for the analysis of petroleum products and lubricants appear in Table 8.

12.3 The calibration standards used in XRF analysis are similar to those used in ICP-AES analysis. Since for petroleum products analysis usually XRF is a solution analysis, the standards are prepared from pure organometallic compounds or organic sulfur compounds in base oil or other suitable organic solvents. Particular attention needs to be paid to the purity of the solvent used. Otherwise a blank correction may be necessary for the contaminants in the solvent.
12.3.1 Working calibration standards can be prepared by careful mass dilution of a certified organic compound with an analyte-free white oil or other suitable base material such as toluene, iso-octane, xylene, etc. The concentration of the unknown samples shall lie within the calibration range that is used.

12.4 Some of the calibration standards suggested and widely used include:
12.4.1 Di-n-ButylSulfide, Thiophene, 2-Methyl Thiophene - Used for sulfur analysis.

12.4.2 Polysulfide Oil - Generally, nonyl polysulfides containing a known percentage of sulfur (as high as 50 m%) diluted in a hydrocarbon matrix. They exhibit excellent physical properties such as low viscosity, low volatility, and durable shelf life while being completely miscible in white oil. The sulfur content of the polysulfide oil concentrate is determined via mass dilution in sulfur-free white oil followed by a direct comparison analysis against NIST (or other primary standard body) reference materials.

12.4.3 Metal Sulfonates, Octoates, Ethylhexanoates, or Cyclohexanebutyrates - Used for metal analysis such as Ba, Ca, Cu, Mg, P, Zn.

12.5 The calibration standards and check samples should be stored in glass bottles in a cool dark place until required. The glass bottles should be either dark or wrapped in opaque material and closed with glass stoppers, inert plastic lined screw caps, or other equally inert, impermeable enclosures. As soon as any sediment or change of concentration is observed, the standard should be discarded.

12.6 The calibration curve is established by determining the net intensity of the emitted radiation from the metal of interest in each of the standards. Usually five to ten calibration standards may be employed; these may be in the mg/kg or 0.x mass% levels depending on the level of the analyte to be determined in the samples. Usually a calibration model is prepared by the software in the XRF instrument.

12.7 Respective X-ray cups are filled at least half full with the calibration standard solutions. No wrinkle or bulges should appear in the film which shall be flat. The filled cups with a vent hole punched in the top of the cell are placed in the X-ray beam in order to measure and record the net intensity (that is, peak intensity - background intensity) for the analyte signal. Although many XRF instruments can count for times even greater than 15 min, in petroleum products analysis a counting period up to 60-s may be used at each wavelength position. This is done for each of the calibration standards for each of the elements measured. A regression analysis is performed for each calibration element by plotting on a linear graph paper or by the way of instrument computer system. It is recommended that a multiple linear regression be performed for each calibration. The regression analysis will determine a slope and an intercept for each calibration element that will be used to determine elemental concentrations in the samples to be analyzed.

12.8 The initial calibration to obtain the slope, intercept, and interelement correlation factors is performed initially when the test method is set up, after any major instrumental maintenance is performed that can affect the calibration (for example, new X-ray tube installed, new crystal added, and so forth), and as deemed necessary by the operator (for example, triggered by quality control sample results). Subsequently re-calibration is performed with a minimum of three standards containing each of the calibration elements at nominal concentrations across the respective calibration ranges in order to check the values of the slope and the intercept.

12.9 Immediately after completing the calibrations, the sulfur concentration of one or more of the calibration check standards is determined. The difference between the two measured values should be within the repeatability of the test method being used. When this is not the case, the stability of the instrument and the repeatability of the sample preparation may be suspect, and corrective measures should be taken. The degree of matrix mismatch should also be considered when evaluating a calibration.

12.10 Drift Correction Monitors - The use of drift correction monitors for determination and correction of instrument drift can be advantageous. Monitors are stable, solid disks or pellets containing all elements of interest in the test method used. Two disks are preferred to correct for both sensitivity and base line drifts. The high-concentration drift monitor provides high-count rates, so that for each analyte, counting error is less than 0.25% relative. The low-concentration drift monitor provides low-count rats, so that for each element, count rate is similar to that obtained with the calibration blank, or zero mass % standard.