Determination of S, Ca, Fe, Ni and V in crude oil by energy dispersive X-ray fluorescence spectrometry using direct sampling on paper substrate
Introduction
Prospected crude oil is a complex matrix [1], [2] composed mainly of hydrocarbons with smaller portions of organic compounds containing S, O and N and organic compounds containing metals [2]. Water and sediments (sand, mud, silt and precipitates) are also present composing the basic sediment and water content. A fraction of water is emulsified within the oil [3].
Among the most abundant metalorganic species in crude oil are the ones of V, Fe and Ni, present in concentrations up to hundreds of μg g−1 [2], [3], [4], [5]. Relative quantities, for instance the Ni/V ratio, bring important information on petroleum formation and thermal maturity, oil migration and accumulation in deposits [3], [5]. Iron can also be incorporated in crude oil due to corrosion of storage tanks and pipelines [6]. Other metals (Ca, Mg) are usually in combination with naphthenic acid as soaps or present either as dissolved or precipitated salts, mostly chlorides and sulfates within the emulsified water [3]. Accessing concentrations of metals in crude oils is also important to indicate potential deleterious effect on catalysts used in oil processing. Sulfur is present in petroleum as organic compounds (mostly as thiophenes), as sulfidric acid and even in its elemental form [3], [4]. The S content requires proper evaluation because of environmental concerns and due to its influence on the acidic properties of the petroleum derivatives [7].
The routine quantification of metallic elements in petroleum is most frequently made by flame atomic absorption spectrometry (FAAS) [2], [6], [8], [9], [10] and inductively coupled plasma optical emission spectrometry (ICP OES) [2], [6], [9], [11], [12]. In special cases, electrothermal atomic absorption spectrometry or inductively coupled plasma mass spectrometry are employed as recently reviewed by Duyck et al. [2], Korn et al. [6], Sanchez et al. [13] and Amorim et al. [9].
The use of methods based on FAAS and ICP OES requires oil sample treatment to get the analytes in aqueous solutions. In other cases, the oil sample is directly diluted in an organic solvent prior to the introduction into the flame or plasma. For instance, the standard guidelines for the determination of metals (Ni, V and Fe) in crude oils by FAAS (ASTM D5863) [10] indicates two possibilities: First, the sample (up to 20 g) is heated to ashes in platinum capsules open to the atmosphere, which takes from 3 (lighter petroleum products) to up to 15 days (for residual oils) to be completed. Then, residual carbon is burnt in a muffle at temperatures above 500 °C. Sample ashes are dissolved in mineral acid prior to the determination. Alternatively, sample is directly diluted in purified kerosene to access the quantities of Ni, V and Fe in the form of metalorganic species but neglecting the analyte fraction within the insoluble particulates. The ASTM D7691 [12], which covers the determination of Fe, Ni, V and other elements, also relies on the introduction of the oil in the (ICP) after dissolution in organic solvents, thus taking into account only the solubilized part of the sample. In the ASTM D5708 [11], oil is reduced to ashes (as indicated in ASTM D5863) prior to the determination of metals by ICP OES. For sulfur, ICP OES determination is impaired as its main spectroscopic lines lies below 190 nm besides low efficiency of excitation and the volatility of its compounds.
The ASTM D5453 [14] covers the total S determination (down to 1 mg kg−1) in liquid hydrocarbons placed into a combustion tube to convert S-containing compounds into SO2 that is excited with UV radiation prior to luminescence detection. The ASTM D1266 [15] is a cumbersome procedure to convert S-containing compounds in petroleum products into free sulfate in aqueous solution prior to gravimetric analysis after precipitation with barium (0.01–0.4 wt.%) or by turbidimetric determination (in the mg kg−1 level).
Despite the fact that the scientific literature indicates the introduction of oil prepared as oil-in-water microemulsions [16], emulsified in surfactant solutions [17] and even in aqueous solution after microwave assisted dissolution [18], these methods are still not recognized in official guidelines due to either the high dilution of samples or due to issues concerning the sample size and the representativeness of the oil.
X-ray fluorescence (XRF) spectrometry requires virtually no sample preparation [5], [19], [20], [21], allowing direct probing of the oil by the X-ray, eliminating interferences caused by differences in physical properties of samples compared to the ones of the standards (affecting transport and nebulization) commonly found in FAAS and ICP OES, due to differences in the composition of standard solutions and sample solutions [20], [21]. However, accurate results in quantitative XRF analysis require strategies to compensate interferences imposed by the sample matrix. For instance, despite the selective capability of wavelength dispersive (WD) XRF spectrometry, the ISO 14597 describes the determination of V and Ni in petroleum products using Mn as internal standard to compensate matrix interferences [22]. The determinations of Ca and S in unused lubrication oil, ASTM D4927 [23], ASTM D6443 [24] and ASTM D6481 [25], by XRF spectrometry takes into account inter-element interferences by using internal standards (Sn in the case of Ca and S), mathematical corrections (requiring information on all elements presents in the sample) or a multi-element calibration matrix with organic elemental standards dissolved in base oil.
For S determinations in petroleum products, the ASTM D2622 [26] employs WDXRF spectrometry, taking into account both the C/H ratio and interferences from other elements in the sample. The ASTM D4294 [27] uses energy dispersive (ED) XRF spectrometry, compensating interferences by spectra deconvolution and mathematical inter-element correction. ISO 14596 indicates the use of Zr as internal standard for S in petroleum products by WDXRF, improving accuracy in determinations [28].
Crude oil and petroleum derivate liquid samples analyzed by XRF (either ED or WD) spectrometry rely on the X-ray radiation directly probing the liquid sample. This is prone to matrix interferences since the penetration of the radiation in the internal layers of the sample results in the absorption of both excitation radiation and fluorescence from analytes in an extent that depends upon the composition of the matrix.
Total reflection XRF (TXRF) spectrometry is a way to minimize such interferences as the radiation only probes the surface layers of the sample. TXRF requires a more complex and expensive instrumentation than EDXRF. Ojeda et al. [5] determined V, Fe, Ni and S in Venezuelan crude oils using TXRF and discussed the advantages and limitations of the technique. Light crude oils produced results above the expected ones probably due to the evaporation of sample caused by heating during sample irradiation. Akpan et al. [29] used TXRF for the elemental trace analysis (Ni, V, Fe, S and many others) of Nigerian crude oils. The samples were diluted in toluene. Sulfur could not be determined because the values were below the detection limit of the instrument.
The sampling of the oil onto solid substrates is a way to enable a sample thin film that imposes minimal matrix interferences. Marguí et al. [30], [31] reviewed the procedures to pre-concentrate liquid samples onto solid substrates prior the analysis by XRF. For instance, Igarashi et al. [32] used liquid–liquid extraction to pre-concentrate metal chelates in an organic solvent that was then deposited on a filter paper before EDXRF analysis. Larner and Seen [33] and Almeida et al. [34] took advantage of the diffusion gradients of metal ions in paper, accumulating them on selective binding agents before detection by EDXRF. A few works are dedicated to petroleum samples, for instance, Teixeira et al. [20] filtered gasoline on cellulose paper prior to EDXRF determination of Fe (down to 10 μg L−1) and Cu (down to 15 μg L−1). Iwasaki and Tanaka [35] determined V, Ni and Fe (down to 0.03 μg) in residual fuel oils by XRF after acid dissolution of sample, complexation of analytes as dimethylcarbamates and pre-concentration in membrane filters. It is important point out that the majority of these works uses the solid substrate mainly to concentrate the analytes present in liquid sample prior to measurements. For crude oil samples, residual fuel or heavy petroleum samples, such analyte concentration is not feasible because of matrix viscosity.
Measured fluorescence intensities depend not only upon the analyte concentrations in the probed area but also on the concentration of concomitant elements in the sample. As X-ray absorption profiles are wide, primary fluorescence are reduced by the absorbing elements in the optical path toward the analyte (pre-filter effect) while the secondary fluorescence might also be absorbed by other elements in the way toward the detector (post-filter effect). Besides, in liquid samples, the film thickness and viscosity also play important role in such matter. In addition, fluorescence from excited matrix elements might also excite the analyte producing unexpected signal reinforcement. In crude oils, interferences of such type can be better handled since the light elements C and H are the main components of sample. Despite their high content, these elements present relatively low mass absorption coefficients. Other representative elements are present in minor quantities. Trace metals with high mass absorption coefficients tend not to impose interferences unless their relative amounts become an issue.
In order to enable prompt quantitative determinations of S (usually present in minor quantities) and Ca, Fe, Ni and V (present in the μg g−1 range) in crude oils, in work, an EDXRF spectrometric method was developed for the direct and simultaneous determination of these analytes after placing either crude oil or vacuum residue samples onto filter paper and using Fundamental Parameter (FP) for quantification with ASTM samples to fit the calibration model. Vacuum residue samples are the ones obtained from the vacuum distillation of the residue from the atmospheric distillation of the crude oil. It consists of hydrocarbons with carbon chains greater than C34 and boiling points above 495 °C).
Section snippets
Materials
In order to develop the FP method, ASTM Proficiency Test samples (F 61009, F 61201, F 61001, F 61005 from ASTM International, USA) were used. During validation of the method, these same ASTM samples in addition to other ASTM Proficiency Test Samples (F 61301, F 61109, F 61101, F 61105, CO 1303, CO 1307 – ASTM International) and a standard reference fuel oil (NIST-SRM 1634c from NIST, USA) were analyzed. Seven crude oils and three vacuum residues were obtained from Petrobras (The Brazilian
Crude oil sample handling and XRF measurements from paper substrates
In order to perform direct analysis of crude oils, enabling reliable quantitative results, efficient sample homogenization is required as indicated in the general recommendations of ASTM D5854 [40] that covers procedures required to ensure representative sampling of petroleum and petroleum products. A study addressing crude oil homogenization has been carefully made in a previous work [41] when proper mechanical agitation of the oil, with an Ultra Turrax device, enabled very long term stability
Conclusions
The determination of trace quantities of Ca, Fe, Ni and V and also S were achieved by EDXRF spectrometry. Samples deposition paper substrate to minimize sample thickness and the inclusion of average C, H, N, O values for matrix correction enabled accurate results for fundamental parameters method with excellent results when response model is obtained with standard crude oil samples. The proposed approach was simple since it requires only a sample dilution to reduce its viscosity before spotting
Acknowledgements
Authors thank the Brazilian Energy Company (Petrobras) and the Brazilian agencies CNPq, FAPERJ and FINEP for financial support. Aucelio thanks CNPq and FAPERJ for scholarships. The skill and dedication of the technicians Leila Fialho and Caio Mello are acknowledged. Authors thank Dr. Eric C. Romani (Physics Department of PUC-Rio) for the profilometric measurements.
References (43)
- et al.
Feasibility for direct rapid energy dispersive X-ray fluorescence (EDXRF) and scattering analysis of complex matrix liquids by partial least squares
Appl Radiat Isot
(2012) - et al.
The determination of trace elements in crude oil and its heavy fractions by atomic spectrometry
Spectrochim Acta, Part B
(2007) - et al.
Determination of V, Fe, Ni and S in petroleum crude oil by total-reflexion X-ray fluorescence
Spectrochim Acta, Part B
(1993) - et al.
Atomic spectrometric methods for the determination of metals and metalloids in automotive fuels – a review
Talanta
(2007) - et al.
Determination of vanadium in petroleum and petroleum products using atomic spectrometric techniques
Talanta
(2007) - et al.
Determination of trace elements in petroleum products by inductively coupled plasma technical: a critical review
Spectrochim Acta Part B
(2013) - et al.
Electrothermal atomic absorption spectrometric method for the determination of vanadium in diesel and asphaltene prepared as detergentless microemulsions
Microchem J
(2004) - et al.
Simultaneous determination of copper and iron in automotive gasoline by X-ray fluorescence after pré-concentration on cellulose paper
Talanta
(2007) - et al.
Determination of sulfur, nickel and vanadium in fuel and residual oils by X-ray fluorescence spectrometry
Anal Chim Acta
(1987) - et al.
Trace and ultratrace analysis of liquid samples by X-ray fluorescence spectrometry
Trends Anal Chem
(2014)