Development of a rapid-test method for the determination of calcium, zinc, phosphorus, and sulfur in automotive engine oil by WD-XRF (wavelength dispersive x-ray fluorescent)

This study describes a rapid method for the determination of calcium, zinc, phosphorus, and sulfur in lubricating oil products using WD-XRF (Wavelength Dispersive X-ray Fluorescent). Currently, there are several normative references appointed by the mandatory SNI (Indonesian standard) used for elemental analysis in automotive engine oils, i.e., ICP-AES (Inductively Coupled Plasma - Atomic Emission Spectroscopy), AAS (Atomic Absorption Spectrophotometry) for calcium, zinc, or magnesium; ED-XRF (Energy Dispersive X-ray Fluorescent) for sulfur, lamp method for sulfur, and quinoline phosphomolybdate method for phosphorus. The wet chemical analysis principle used in those reference methods (AAS, ICP-AES, and conventional method) needs a sample preparation step before the main instrument can use to measure an oil sample. It is a time-consuming process and commonly consists of a few steps such as dilution, weighing, heating, and destruction. The statistical characteristics (limit of detection, the limit of quantification, precision, accuracy, and linearity) obtained in this experiment were relatively acceptable. Elemental analysis by WD-XRF needs no sample preparation process, fast, less chemical reagents, multi-elements, simultaneously, and simple. The WD-XRF method potentially could be used as an alternative method for the determination of certain elements (Ca, Zn, P, S) or more as specified in the mandatory SNI for automotive engine oil products.


Introduction
Lubrication is one of the critical factors in maintaining engine performance in optimum conditions. The lubricating oil can protect an engine from degradation factors such as wear, heat, corrosion, oxidation, and various contaminant [1]. The quality of oil used in engine parts will affect fuel consumption and exhaust gas. Optimum lubrication can reduce gas emission produced by a combustion engine [2]. Good lubricating oil could preserve engine performance, cost efficiency, less air pollution, and increase safety.
The number of motor vehicles that continues to grow every year will boost the need for lubricants in Indonesia. According to BPS (Statistics Indonesia) data, in 2018 the total number of four types of motorized vehicles (passenger cars, bus cars, freight cars, and motorbikes) reached 146,858,759 vehicles [3]. The need for lubricants for such large numbers of motor vehicles needs to be balanced by reliable lubricant quality.
The specific element's content is included in the mandatory SNI specifications. The elements that should be determined in lubricating oil products are calcium, magnesium, zinc, phosphorus, and sulfur. According to the SNI, each element could be analyzed by a few different methods. Table 1 shows the reference methods used in the SNI. Metals content in lubricating oil provided valuable information about its composition, sedimentation of metal parts, or degradation of additive in an engine [4]. Elements contained in automotive lubricating oils were found as impurities or additives. Calcium and zinc in automotive lubricating oil were used as a detergent additive ingredient and anticorrosion agent [5]. Phosphorus in engine oil was seen as anti-wear/antioxidant additive that, during operation, could be volatilized [6]. Sulfur in engine oil was found as mineral base oil impurities [7]. Heavy metals (Al, B, Cr, Cu, Fe, Pb, Mn, Mo, Ni, Na, Si, Ag, Sn, Ti, V, Cd, and As) in used oil can provide valuable information about wear metals and contaminants [8]. The Determination methods of elements content in lubricating oil according to ASTM standards are as follows: (a) Calcium, magnesium, and zinc; tested by AAS (Atomic Absorption Spectrophotometry) [9] or ICP-AES (Inductively Coupled Plasma -Atomic Emission Spectroscopy) [10]; (b) Phosphorus tested by quinoline phosphomolybdate method (titration) [11] or ICP-AES [12]; and (c) Sulfur tested by lamp-method [13], WD-XRF [14], ICP-AES [12] or ED-XRF [15,16]. The wet chemical analysis technique was used in AAS, ICP-AES, and titration.
Elemental analysis in lubricating oil product using AAS method has a good detection limit [17]. Still, it needs a time-consuming sample preparation process, and cannot be determined multi-elements simultaneously. For ICP-AES, although it has a very low detection limit and could measure multielements at once [8], it needs some expensive consumable supporting materials and precise sample preparation process. Titration method for phosphorus determination by quinoline phosphomolybdate is a low-cost method (no need an expensive instrument like AAS, ICP-AES, nor XRF), but it needs a complicated preparation process and more chemicals. Among the other available analytical techniques, XRF spectrometry potentially can be used as a single method for determing all of the specific elements required by the mandatory SNI specification simultaneously. WD-XRF (Wavelength Diffraction X-Ray Fluorescence) method has developed for the analysis of Al, Ba, Ca, Cr, Cu, Fe, Mg, Mn, Ni, P, Pb, S, Si, Ti, V, and Zn in various petroleum liquid products with no significant loss of accuracy [18]. The WD-XRF method offer advantages by its minimum preparation, fast, and has an excellent detection limit compared to other elemental analysis techniques.
This study focuses on determining the concentration of selected elements (Ca, Zn, P, and S) in automotive engine oils by using the WD-XRF method for described a simple and fast testing method. Statistical parameters for method validation is also presented [19].

Materials
ASI (Analytical Services Incorporation) LOE17 was used as Oil Reference Material. Commercial, automotive engine oil that available in markets in Bandung region was used as a lubricating oil sample. Plastic X-ray cell of 25 mm diameter and a plastic film chemically resistance (Mylar®, with 6 μm thickness) was used as a base holder material.

Instrument
ASI (Analytical Services Incorporation) LOE17 was used as Oil Reference Material. Commercial, automotive engine oil that available in markets in Bandung region was used as a lubricating oil sample. Plastic X-ray cell of 25 mm diameter and a plastic film chemically resistance (Mylar®, with 6 μm thickness) was used as a base holder material.

Procedures
About 50 mL standard or sample was added to a cell cup then placed in an autosampler holder. The measurement of metals content was performed by WD-XRF Analyzer and controlled by PC. The calibration standards were prepared from CRMs (Certified Reference Materials) diluted by white oil; at least five standards solution were measured to obtain the calibrations curves. The measurement of samples was carried out after the calibration curves were made. Every sample was measured at least three times (triple) to have a representative result. The Limit of Detection (LD) and the Limit of Quantification (LQ) value was determined by measuring a new solution (white oil) or spiked sample for five times. Measured CRM determined the Accuracy and Precision for five times, then compared the value obtained from the measurement with the actual number stated in the certificate. These experiments were evaluated using CRM as a control to ensure the accuracy of the analysis.

Time of analysis
A simple procedure was applied in this experiment. Lubricating oil sample from the original sealed bottle poured into a cell cup up to volume marker (about 50 ml), then it placed in autosampler holder. The measurement proceeds directly after the cell cup placed in the autosampler holder. The whole process for conducting elemental analysis for a single sample took about 3-5 minutes or less than 5 minutes.
The common elemental analysis method consisted of three necessary steps: sample preparation, measurement, and calculation. The modern analytical instrument commonly had equipped by software that could be used for calculating the instrument data to generate the measurement result, so the count took place during the measurement process. The main issues in the analysis duration were the sample preparation process. The wet chemical sample preparation commonly needed a few procedures, such as weighting, dilutions, and digestion. This process could become an uncertainty contributor and affected the accuracy of the method. The time needed during a sample preparation process (i.e. weighthing, dissolved, making standard solution, and dilution) was varied, depended on the sample characteristics (i.e. solubility, and suspended solid) and instruments (i.e gas type, light source, and compatibility) used. A fast analytical method or rapid-test would have a lot of advantages if applied in-field inspection and quality control processes in the factory. Less chemical means more contribution to green industrial production. Time efficiency would be to reduce the cost of analysis and human resources.

Limit of detection and limit of quantification
Limit of Detection (LD) or Limit of Quantification (LQ) was commonly defined as the lowest concentration of an analyte that can be detected. Several approaches can be used to estimate LD and LQ. i.e., visual assessment, signal to noise ratio, calculated from the standard deviation of the blank, calculated from the calibration line at low concentration [20]. LD was meant to know the limit test for impurities, and the LQ was to determine the quantitative analysis for the impurities quantification of the main component [21].
The LD and LQ were calculated from the measurement of a spiked sample. The sample was measured seven times under reproducibility conditions. It was calculated by concentration corresponding to the blank or spiked sample plus three standard deviations, as shown in equation 1, and LQ is corresponding to ten standard deviations, as shown in equation 2 [20].
(1) ( = + 10 (2) where x is the mean concentration of the blank or spiked sample, and σ is the value of the standard deviation of the measurements. Although by this method it was difficult to prove that a low concentration of analyte signal was different from the blank signal, it had the benefit as the simplest and fast method that could be used in LD and LQ determination. Table 2 shows that LD (%mass) for Ca = 0.00955, P = 0.00800, S = 0.06907, and Zn = 0.03075; and LQ (%mass) for Ca = 0.01693, P = 0.01277, S = 0.07508, and Zn = 0.03442.

Precision and accuracy
The precision of the method was calculated by analyzing a series of samples from a homogenous solution. The Collaborative International Pesticides Analytical Council Limited (CIPAC) guidelines stated that the repeatability precision would be acceptable if several replicated samples were not less than five replicates and the value of %RSD calculated from measurement less than CV-Horwitz multiple by 0.67 [22]. Data from table 2 shows that the %RSD of analysis for Ca, P, S, and Zn were below the 0,67 x CV-Horwitz value. %RSD was calculated by equation 3 Where, x is measurement results, σ is standard deviation of measurement, and C is measurement results in percent (%). The measurement result generated the accuracy of the testing method to the actual value stated in CRMs. The accuracy was evaluated by using z-scores (ⱬ), a statistical parameter that indicates how many standard deviations of the measurement results compared with the certified/true value of an element in CRMs, z-scores was calculated by the equation 5. The interpretation of the z-score is satisfactory if the score not more than 2 (|score|≤2), unsatisfactory if more than 3 (|score|≥3), and questionable if the score between 2 and 3 (2<|score|<3) [23]. According to table 3, the accuracy of the measurement of sulfur, phosphorus, and zinc was satisfied, but for calcium, it needed to be improved.

Calibration curve and linearity
The linearity of response to the analyte concentration should be calculated at least three measurement data. The curved showed no significant deviation from linearity if the correlation coefficient (R2) value was above 0.99 [22]. Interferences and sample matrix need to consider in constructing a working range because obstructions can cause non-linear responses, and sample matrix could change the ability of the method to extract/recover the analyte from the sample. The working range and linearity for this method were assessed by visual inspection of plotted data in the calibration curve supported by statistical linear regression. The calibration curve needed to be established individually for each element and matrix to get a quantitative analysis. Working range proposed for a sulfur determination as shown in figure 1 was between 0.002 to 5 mg/kg with linearity coefficient (R 2 ) value 0.99713, for calcium shown in figure 2 was between 0.0025 to 0.5 mg/kg with linearity coefficient (R 2 ) value 0.99964, for zinc shown in figure 3 was between 0.05 to 0.25 mg/kg with linearity coefficient (R 2 ) value 0.99643, and for phosphorus shown in figure 4 was between 0.002 to 5 mg/kg with linearity coefficient (R 2 ) value 0.99713.

Uncertainty of measurement
Sources of uncertainty in XRF spectrometry according to IAEA (International Atomic Energy Agency) consists of six components, i.e., calibration of the spectrometer, instability of spectrometer (detector and electronics), sample preparation (heterogeneity of the material and non-uniformity of sample thickness), spectral data processing with the fitting program, quantification (determination of the absorption correction factor), and uncertainty in the determination of total mass per unit area [24]. The uncertainty contributors in XRF measurement were influenced by the calibration curve, variation of replicates, sample preparation, and influence from operators. Uncertainty contributors from calibration measurements and laboratory analysts were negligible [25].
The estimation of uncertainty is simple in principle. The following approached could use to the quantification of uncertainty associated with a measurement result in different circumstances, i.e., using data from in-house and collaborative method validation studies, Quality Control, Proficiency Testing, relevant prior studies, and the use of formal uncertainty propagation principles [26]. Uncertainty estimation for elemental analysis of lubricating oil by WD-XRF base on the testing process indicated that the best available contributor was from the individual repeatability of the measurement.

Qualitative and quantitative measurement
WD-XRF is a powerful tool in elemental analysis problems. The method is proposed for qualitative and quantitative analysis of solid material (thin or bulk, alloys or rocks, etc.), with advantages in sample preparation (without treatment, fusion, etc.) and measurement (single or multi-element) [27]. Qualitative and quantitative analysis can be used for the determination of major, minor, and trace elements in various kinds of samples. The qualitative analysis (scanning mode) can be done satisfactorily for automotive engine oil samples, for quantitative analysis depends on the availability of oil standards (CRMs, SRM). Table 4 shows the analysis results for various types of automotive engine oils. The twelve samples sequentially analyzed by WD-XRF and finished in 50 minutes.

Conclusion
WD-XRF method allows the simultaneous determination of certain elements (calcium, zinc, phosphorus, and sulfur) in lubricating oils quantitatively. Complicated sample preparation was not needed, and it was simple in terms of its analysis procedures (no need to burn the sample, heated,