Elemental analysis as an important tool for fuels and biofuels quality control

Article presents normalized analytical methods for determination of elements in fuels and biofuels with emphasis on inductively coupled plasma emission spectrometry (ICP-OES) and their advantages and disadvantages. Analytical procedure applied for determination of 20 elements (Al, B, Ba, Ca, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, Pb, Si, Sn, Ti, V and Zn) is described with elements of validation. Broad range of elemental analysis results of fuels (B0, B7 and B10) and pure FAME samples are described. Typical contents of selected elements are presented which allows discussion on the domestic fuel quality in regard to non-standardized parameters and their poten-tial effect on combustion engines performance.


Introduction
Analysis of liquid petroleum products characteristics is a challenge for every chemist due to variety of these products (petroleum, diesel fuel, biofuels, engine oils, gear oils, hydraulic oils, coolants etc.). Conclusion regarding these products quality are usually based on numerous laboratory analyses.
In order to determine diesel fuel quality analyses are required such as cetane number and cetane index, distillation characteristics, density, viscosity, fatty acid methyl esters (FAME) content, water content, polycyclic aromatic hydrocarbons content, flash point, carbon residue, total contamination, copper strip corrosion, oxidation stability, lubricity and climate dependent cold filter plugging point and cloud point. Within the analyzed range of properties are also concentration of manganese and sulfur which content is checked and verified against EN 590 standard's requirements, as the only elements in diesel fuel which concentration need to be indicated directly [1]. In case of petrol concentration of three elements need to be verified against EN 228 [2] requirements namely: lead, sulfur and manganese. Other parameters which has to be checked to control fuel quality are: distillation characteristics, vapour pressure, density, research and motor octane number, oxidation stability, existent gum content, copper strip corrosion, aromatics and olefins hydrocarbons content, oxygen and oxygenates content and benzene content. More detailed elements analysis is required in case of fatty acids methyl esters [3], in which sulfur, phosphorus, sodium, potassium, calcium and magnesium concentrations need to be checked in order to verify this component quality. In fact, normative requirements regarding these elements concentration are only ones allowing estimation of some guideline values for discussion on elemental concentration in regular fuels. These should be free from any adulterant or contaminant that may render the fuel unacceptable for use in engine vehicles as it is described in both EN 228 and EN 590 standards [1,2].
Determination of concentration of wide variety of elements requires proper sample preparation before measurement, as well as application of suitable measurement technique which would allow reaching very low limits of determination (LOD) leading to precise and repeatable results despite the demanding organic matrix. This paper discusses trace elements analysis in liquid fuels: diesel and biodiesel fuels. In BOSMAL Automotive Research and Development Institute Ltd fuel quality check includes routine determination of 20 elements (Al, B, Ba, Ca, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, Pb, Si, Sn, Ti, V and Zn) concentration in case of diesel and biodiesel. In case of fuels without biocomponents concentration of these elements are usually below 1 mg/kg.

Legal requirements
As it was mentioned above, only limited number of elements is an object of normative regulation. In case of sulfur and manganese 10 and 2 mg/kg, respectively, is upper allowed concentration in petrol and diesel, and additionaly in case of petrol lead content is limited to max 5 mg/kg. In case of FAME used as a 100 % biofuel or blending component in diesel fuel, six elements content is limited namely: sulfur (max 10 mg/kg), phosphorous (max 4 mg/kg), summary content of sodium and potassium content limited to max 5 mg/kg, exactly the same as summary content of calcium and magnesium. On the other hand sodium, zinc and silicon content above 2 mg/kg may have negative impact on injectors performance of modern diesel engines, furthermore increased concentration of some elements may result in engine components corrosion and formation of deposits at elevated temperatures [4,5]. Taking into account these negative aspects of trace elements presence, their concentrations in fuels should be monitored, especially during the demanding process of engine testing.

Methods of trace elements determination
Trace elements concentration is usually measured by spectroscopic methods, mainly inductively coupled plasma optical/atomic emission spectrometry (ICP-OES, ICP-AES) and wavelength or energy dispersive X-ray fluorescence (WD-XRF, ED-XRF). Comparison of standardized methods with description of sample preparation, measurement parameters for different petroleum products and concentration ranges is presented in Table 1 and 2.

Sample preparation for ICP-OES/AES measurements
One of most important stages, in each analytical procedure is sample preparation. Improper sample preparation may lead to enrichment or impoverishment of analitu, for example due to contamination. That is why the sample preparation procedure should be as simple as possible to minimalize possibility of error. Sampled fuel should be homogenous and usually requires placing in water bath [6], homogenization by intensive mixing [7-10, 14, 21], heating up to 60-100ºC and intensive shaking [4,11] or by mechanical homogenization [14].
After homogenization, in case of ICP-OES or ICP-AES, usually sample is diluted with pure solvent with viscosity and boiling point similar to the sample. Alternatively direct measurements of undiluted samples are also possible. However it is important to bear in mind that both of this approaches are improper if insoluble particles are to be taken into account in elemental analysis, and may lead to underestimation of results [4,14]. It is of high importance that standards solutions and samples have very similar viscosity as this is the critical parameter influencing spraying efficiency in spray chamber of spectrometer. That is why both, samples and standards solutions, should be prepared with the use of the same solvent. Additionally, in order to increase accuracy of the results internal standards (IS) are used and their role is to correct matrix effects and slight differences between viscosities of the samples and standards solutions. Internal standards are used in form of metalorganic compounds of usually: Sc, Y or Co, which are not present in the sample. In case of highly volatile samples as ethanol or petrol measurements may be performed on undiluted samples however, it may result in unstable plasma work and cooled spray chamber may be required.

Experimental
Elements concentration was measured by inductively coupled plasma optical emission spectrometer Perkin Elmer Optima 8300 with horizontal plasma and binary, radial and axial, observation system.
Due to low concentration of elements, below 1 mg/kg, samples after homogenization were analyzed undiluted. International standards recommends minimum 10x dilution what significantly deteriorate determination limits which in result makes accurate determination of trace elements difficult. Such approach is relatively easy for petrol and diesel containing up to 10%vol. of FAME (B10), however in case of pure FAME its viscosity is too high and sample dilution was required. For dilution of high viscosity samples V-solv (VHG Labs, Manchester, USA) solvent was used (declared concentration of elements below 0.25 µg/g).
Before measurements optimization of spectrometer parameters was done for each analyzed element, in order to achieve limit of determination as low as possible, signal to background ratio as high as possible, repeatable measurements and stable plasma work. Selected working conditions are presented in Table 3. For each analyzed element plasma gas flowrate, auxiliary gas flowrate, nebulizer gas flowrate, plasma power and observation direction were optimized. For most of analyzed elements, except Al, B, Cr, K and Na, 2 or 3 analytical lines were chosen and applied. In case of potassium and phosphorous binary plasma observation, radial and axial, was applied. Obtained limits of quantification (LOQ) were: 0.1 mg/kg for Al, Ba, Ca, Cu, Fe, Mg, Mn, Pb, Si, Ti, V and Zn; 0.3 mg/kg for Cr and 0.5 mg/kg for B, K, Mo, Na, Ni, P and Sn.
Four point calibration curves were prepared for each analyzed element and were based on multielemental standard solutions obtained by dilution of basic standard solution of elemnts in metalorganic form and concentration of 100 mg/kg each (WM-22-10x-100, AccuStandard, New Heaven, USA). Correlation coefficient for most elements was above 0,9995, except: Na (r 2 = 0.9991) and P (r 2 = 0.9979). Correctness of developed method was checked by testing of measurement repeatability (relative variation coefficient for 1.0 mg/kg concentration level was from 2.1 to 11.4% depending on measured element), recovery by standard addition method (90.2-114.4%, except B: 80.4%) and correctness of calibration curves preparation (98.7-108.4%).

Results and discussion
The aim of conducted research was quality check of fuels from Polish market. Over 370 samples of fuels: diesel without FAME (B0), diesel with up to 7 %vol. FAME (B7), diesel with up to 10 %vol. FAME (B10) and FAME (B100) were tested within four years between 2013 and 2017. Within these samples there were 96 samples of B0, 207 samples of B7, 66 samples of B10 and 10 samples of B100.
Average elements contents in tested fuels are presented in Figs 1-3. Comparison of measured elements concentrations including minimum and maximum values is presented in Table 4, and these are the typical elements contents in tested fuels. Elements most frequently present in B0, B7 and B10 fuel samples were: Al, Ca, Cu, Na, P, Pb, Si and Zn. Within these elements sodium was predominant taking into account both: its concentration in fuels and its presence frequency. Average sodium content in B0, B7 and B10 was ca. 1,0 mg/kg, and it was present in 90% of tested B0 samples and 78% of B7 and B10 fuel samples. In case of pure FAME, B100 samples sodium content was significantly higher (6.3 ± 1.3 mg/kg), however only 10 samples of B100 were tested. In case of silicon, average concentration in B0 samples was 0.5 mg/kg with maximum of 1.6 mg/kg (presence in 81% of B0 samples), in B7 average concentration was 0.3 mg/kg with maximum at 1.0 mg/kg (presence in 57% of samples) and in case of B10 average content of silicon was 0.2 mg/kg with maximum of 0.6 mg/kg (presence in 68% analyzed samples). It was noticed that average silicon content was decreasing with increasing of FAME content in diesel fuel. Zinc was present in tested fuels samples at average level of 0.22 ±0.13 mg/kg and no relation to biocomponent content in fuel was noticed. In case of pure FAME average concentration of this element was 0.76 mg/kg.
The highest phosphorous content was noticed in FAME samples (average of 2.8 mg/kg) while in fuels samples it was on average level of 0.5 mg/kg, in the range of minimum 0.1 and maximum 1.7 mg/kg. Similar dispersion of results was observed in case of silicon, average value of 0.3 mg/kg with dispersion of results from 0.1 to 1.6 mg/kg and in case of lead, average value of 0.3 mg/kg with dispersion from 0.1 to 1.4 mg/kg. Phosphorous presence was noticed in case of 31% of B0 samples, 51% of B7 samples and 62% of B10 samples, while lead was present in 44% of B0 samples, 41% of B7 and 32% of B10 samples.
There was no B, Ba, Cr, K, Mn, Sn, Ti, and V presence noticed in case of any of tested samples. Other elements than these mentioned above were present in relatively small number of samples, like magnesium, which presence was noticed only in 4 from 90 tested B0 samples. Moreover, its concentration was 0.1 mg/kg which was also the limit of quantification. Iron and nickel were present in some individual cases ie. Fe was present in 12 samples of B0 and only in 4 samples of B7, while Ni in 5 samples of B7 and 2 samples of B10. There were no iron presence noticed in any of B10 and B100 samples as well as no Ni presence in B0 and B100 samples. Molybdenum concentration in most of the samples was below limit of quantification (0.5 mg/kg) and was estimated in the range of 0.15-0.30 mg/kg. Molybdenum concentration was higher than LOQ only in three cases of B7 fuels. The lowest results dispersion was observed in case of calcium. This element was present in 8% of total samples number and its average concentration was 0.13 ±0.03 mg/kg. Copper, which was present in 26% of total samples, results dispersion was significantly higher with average concentration for this element 0.22 ±0.16 mg/kg.

Conclusions
Negative effect of some elements presence, especially Zn, Na and Si, on engine performance is well known [5]. In some cases relatively small increase in some elements concentration may lead to significant engine performance drop. This is particularly true in case of soap formatting elements like Na, which responsibility for injector cooking is of great importance. The effects of other elements than mentioned above on engine performance is not that well known, mainly due to their trace level presence in most of diesel fuels. However, while dynamic development of combustion engines is taken into account with the accompanying development of alternative fuels production processes, it is possible that some of the elements which presence is not of that great importance, like the ones mentioned earlier, may become more important in the future.
Inductively coupled plasma emission spectrometry is a well-known and mature technique allowing measurements of broad range elements in the limits usually do not achievable for other spectrometry techniques, like X-Ray fluorescence or atomic absorption spectrometry. Different possibilities relating to sample preparation and control of versatile measurement parameters by ICP-OES techniques allows reaching very low LOQ values what is necessary for monitoring of trace elements. Despite the fact that, in case of diesel fuels, only sulfur and manganese content is limited by the standards there is need to other elements monitoring, especially in case of engine testing or vehicle testing activities as their presence may have significant effect on engine performance. The effect of fuel contamination with trace elements in regular car usage will have presumably less dy-namic nature and the engine performance drop will be less noticeable.

AAS
atomic absorption spectrometry ED-XRF energy disppersive X-Ray fluorescence IS internal standard ICP-OES inductively coupled plasma optical emission spectrommetry ICP-AES inductively plasma coupled atomic emission spectrometry LOQ limit of quantification LOD limit of determiantion UV-F ultra-violet fluorescence WD-XRF wavelenght disppersive X-Ray fluorescence Bibliography