- Email: [email protected]
Determination of ethanol in gasoline by high-performance liquid chromatography
Fuel 212 (2018) 236–239
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Full Length Article
Determination of ethanol in gasoline by high-performance liquid chromatography
MARK
⁎
Lorena Morine Avilaa, Amanda Pereira Franco dos Santosb, , Danielle Ignácio Mançano de Mattosb, Cristiane Gimenes de Souzab, Débora França de Andradec, Luiz Antonio d'Avilab a b c
Instituto de Química, Universidade Federal Fluminense, Niterói, Rio de Janeiro 24020-141, Rio de Janeiro, Brazil Engineering Program in Chemical and Biochemical Processes, Escola de Química, Universidade Federal do Rio de Janeiro, 21941-909 Rio de Janeiro, Brazil Instituto de Química, Universidade Federal do Rio de Janeiro, Cidade Universitária, 21941-909 Rio de Janeiro, Brazil
A R T I C L E I N F O
A B S T R A C T
Keywords: Gasoline Ethanol HPLC
This study employed non-aqueous reversed-phase high-performance liquid chromatography (HPLC) with refractive-index detection and methanol as a mobile phase to quantify the ethanol content in gasoline. The advantages of HPLC are its good separation versatility, high resolution, relatively short analysis time, and automation. Standard samples were prepared to obtain a standard curve and partial validation. Accuracy, precision, linearity, selectivity for methanol (adulterant), and measurement uncertainty were investigated. In the partial validation analysis, the accuracy of the proposed method was identified as its main advantage over the reference method.
1. Introduction When fuel is adulterated, it is usually done by adding a lower cost product in order to obtain illicit financial gains from its retail. The illegal addition of excess ethanol to gasoline is arguably the easiest and most common form of adulteration, since ethanol is already a component of the gasoline/ethanol blend used in Brazil (27% ± 1% by volume) and is cheaper than gasoline. According to the national bulletin of the Fuel Quality Monitoring Program run by the Brazilian fuel regulatory agency (ANP), the ethanol content of gasoline is the biggest cause of noncompliance found in samples of gasoline [1,2]. The reference method for determining ethanol content, described in ANP resolution N°40 of October 25, 2013, is detailed in Brazilian standard NBR 13992/2015, issued by the Brazilian technical standards association (Associação Brasileira de Normas Técnicas, ABNT) [3,4]. It is a quick, simple, practical method that can be done in the field to check gasoline purity. However, because of these very features, its measurement uncertainty is 1 vol%, and it has the added limitation of quantifying other water-soluble alcohols like methanol, impairing its accuracy. The adulteration of ethanol by methanol has been verified in Brazil by the ANP [5–8], and this has motivated the development of tests to determine methanol content in ethanol [9,10].
The American Society for Testing and Materials (ASTM) has its own standard, D5501, published in 2012 and reapproved in 2016, which describes a standard method using gas chromatography (GC) for the determination of the ethanol (20 mass% or over) and methanol (0.01 mass%–0.6 mass%) content of fuels [3,11]. This method is not so easy to implement, as it is done using long columns (100 m and 150 m), which, while they do exist, are harder to find on the market and very expensive. In the case of the 100 m column, the initial temperature of the column is supposed to be 15 °C, but this is unfeasible in much of Brazil as it is lower than ambient temperature, meaning that a cooling stage would first have to be employed. Further, the specific density of all the samples must be calculated to correct the results obtained [11]. In Brazil, ABNT set up a working group at its Ethanol Fuel Study Commission to create a reference method for quantifying ethanol and methanol in gasoline and ethanol fuel. The chosen method was GC, but after much research no reference method for quantifying ethanol in gasoline was produced, since the compounds in gasoline are eluted in the same time as ethanol [12]. There are different methods described in the literature for analyzing oxygenates in gasoline for quality control and regulatory purposes, like Fourier-transform near-infrared spectroscopy [13] with partial least square (PLS) calibration [14,15]; Raman spectroscopy [16]; synchronous fluorescence spectroscopy with principal component regression or
⁎
Corresponding author at: Centro de Tecnologia, Bloco K02, LABCOM, Cidade Universitária, Rio de Janeiro 21941-909, Brazil. E-mail addresses: [email protected] (L.M. Avila), [email protected] (A.P.F. dos Santos), [email protected] (D.I.M. de Mattos), [email protected] (C.G. de Souza), [email protected] (D.F. de Andrade), [email protected] (L.A. d'Avila). http://dx.doi.org/10.1016/j.fuel.2017.10.039 Received 21 February 2017; Received in revised form 22 September 2017; Accepted 6 October 2017 0016-2361/ © 2017 Elsevier Ltd. All rights reserved.
Fuel 212 (2018) 236–239
L.M. Avila et al.
Calibrated glassware (10.00 mL volumetric flask, 1.00 mL volumetric pipette, and 10–100 µL variable volume automatic micropipette) was also used in the preparation of this sample.
PLS calibration models [17]; a combination of excitation-emission matrix fluorescence spectroscopy with multiway partial least square regression (N-PLS) and unfolded PLS [17]; batch injection analysis with detection by multiple-pulse amperometry [18]; Terahertz spectroscopy [19]; proton nuclear magnetic resonance spectroscopy (1H NMR) [20]; cyclic voltammetry and multivariate calibration [21,22]; and customized mobile near-infrared spectrometry [23]. ASTM D5599/2010 describes a procedure for determining oxygenate content in gasoline by GC, but only for the 0.1–20% by mass range, which is not enough for the Brazilian fuel market [24]. ASTM D4815/2013 describes a GC method for determining methyl tert-butyl ether (MTBE), ethyl tert-butyl ether (ETBE), diisopropyl ether (DIPE), tert-amyl alcohol, and C1–C4 alcohols in gasoline. However, this method is employed to determine alcohol levels of 0.2 mass% to 12 mass%, ruling out its use for ethanol-based fuels and making it unfeasible for testing Brazilian gasolines [25]. In this context, this study proposes the use of high-performance liquid chromatography (HPLC) with refractive index detection to develop a quick method for quantifying the ethanol content in gasoline. It is hoped that this method may replace the existing reference test for laboratory analyses and the monitoring and inspection of gasoline sold in Brazil, whose drawbacks are its minimum uncertainty of ± 1% (by volume) and the fact that it is not accurate enough for determining ethanol levels in gasoline if it is adulterated with methanol.
2.2. Analysis of the samples by high-performance liquid chromatography (HPLC) All the samples were analyzed using non-aqueous reversed-phase high-performance liquid chromatography (HPLC) with a Dionex UltiMate™ 3000 quaternary pump (Thermo Scientific, Massachusetts, USA), a Shodex RI-101 refractive index detector (ECOM, Czech Republic), and an UltiMate™ WPS-3000 autosampler (Thermo Scientific) with a 100 µL sample loop. An Acclaim™ column (Thermo Scientific) measuring 250 mm long and 4.6 mm internal diameter was used with an octadecylsilane phase with 5 µm particle size and 120 Å pore diameter. The mobile phase was 100% pre-filtered HPLC-grade methanol (Tedia Brazil, Rio de Janeiro, Brazil). Total analysis time was 15 min at 40 °C (temperature of column oven), with a constant flow rate of 1 mL/min and 10 µL injection volume. The samples were injected in quadruplicate to evaluate repeatability. The chromatograms were analyzed using Chromeleon 6.80 SR11 software (Thermo Scientific) with manual integration. 2.3. Partial validation of proposed method In order to use objective evidence to ascertain whether the proposed method produces reliable results that are fit for purpose – i.e. whether it meets the requirements and standards for the specific proposed use – we assessed its linearity, selectivity for methanol (adulterant), accuracy, precision, and measurement uncertainty [26–30]. Linearity was checked by the coefficient of determination (R2) of the straight line. Values over 0.99 indicated the linear working range, where the angular coefficient of the straight line could be considered constant [26]. The selectivity of the HPLC method for methanol was checked by injecting a sample of pure gasoline (matrix) and a standard sample containing 12.5 vol% ethanol, 12.5 vol% methanol, and 75 vol% gasoline, since no certified reference material exists for such a determination. In order to assess the accuracy of the verification samples, we used Eq. (1) to ascertain relative error (Erel) [26]:
2. Materials and methods 2.1. Preparation of standard samples of gasoline with ethanol Eighteen standard samples of gasoline containing different quantities of anhydrous ethanol fuel, ranging from 19 to 40% (by volume), were prepared. The gasoline was supplied by Companhia Brasileira de Petróleo Ipiranga and the ANP inspection department, and support was provided by the Fuel and Petroleum Products Laboratory at the School of Chemistry, Federal University of Rio de Janeiro (LABCOM/EQ/ UFRJ). All the glassware used in preparing the samples (1.00, 2.00, 5.00, and 10.00 mL volumetric pipettes; 10–100 µL variable volume automatic micropipette; and 50.00 mL volumetric flask) were calibrated in advance. Six of the 18 samples were used to plot an analytical curve and 12 were used to verify the analytical curve and to partially validate the proposed method, as shown in Table 1. Aside from the 18 standard samples, a sample of gasoline containing 12.5 vol% ethanol and 12.5 vol% methanol was prepared to verify accuracy when ethanol was adulterated with methanol and to demonstrate its selectivity for ethanol.
Erel =
Standard Samples for Verification
Sample Code
Nominal Ethanol Value (% volume)*
Sample Code
Nominal Ethanol Value (% volume)*
A1 A2 A3 A4 A5 A6
20.00 22.00 24.00 26.00 28.00 30.00
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12
19.00 20.00 21.00 23.00 23.00 25.00 26.00 27.00 27.00 31.00 32.00 40.00
(1)
where: Erel is relative error (%); Xlab is the single value obtained experimentally or the mean of the laboratory results; Xv is the value accepted as being true.
Table 1 Standard samples of gasoline containing different proportions of anhydrous ethanol fuel. Standard Samples for Analytical Curve
xlab−x v ∗100 xv
In this study, the relative error of the 12 verification samples (V1–V12) was calculated, assuming a 95% confidence interval for the statistical treatment of the results obtained. As such, the maximum admissible error was set at 5%. The repeatability limit (r) calculated by Eq. (2) and the total amplitude of the measurements (difference between the highest and lowest value from a dataset) were used to evaluate the precision of the proposed method [27].
r = 2.8∗Sr
(2)
where: r is the repeatability limit; Sr is the standard deviation of repeatability, meaning the standard deviation of the results for each concentration.
* The nominal values were corrected using the data from the glassware calibration certificates (see Table 2, in Results).
237
Fuel 212 (2018) 236–239
L.M. Avila et al.
Finally, the expanded uncertainty of measurement of the HPLC method described here, for a 95% confidence interval, was calculated according to the Guide to the Expression of Uncertainty in Measurement (ISO GUM) [30]. It considered the following sources of uncertainty: the calibration of the pipettes and flasks used to prepare each standard sample; the purity of the ethanol used; the straight-line equation obtained; the preparation of the control sample; and the repeatability of the measurements. For each source of uncertainty, we ascertained what type of uncertainty was involved and what method was best suited to determining and calculating it. We carried out type A and type B evaluations of uncertainty and closely observed the contributions made by more than one source of uncertainty so that no calculations were duplicated. Electronic spreadsheets were used for the calculations of combined uncertainty, effective number of degrees of freedom, coverage factor, and expanded uncertainty. The spreadsheet showing the uncertainty of measurement calculations for this HPLC method is included in the Supplementary material.
Height (μRIU)
area (μRIU*min)
600 400 200 0 12
14
27
29
31
The proposed high-performance liquid chromatography method proved effective for quantifying the anhydrous ethanol fuel content of gasoline for its speed, simplicity, accuracy, and precision, demonstrated by the statistical tools used in its partial validation. However, its greatest advantage is the accuracy it was found to have, as it suffered no interference from the other constituent parts of the matrix (gasoline) or the interferent (methanol). This is one of the drawbacks of the traditional method and the main reason why this research was done. We conclude that the HPLC method described in this study is promising for the accurate determination of the ethanol content of gasoline. The expanded uncertainty of this method (0.56 vol%) is also lower than that of the current reference test (1.00 vol%).
800
10
25
4. Conclusion
Ethanol
8
23
containing 12.5 vol% ethanol and 12.5 vol% methanol in gasoline A, highlighting the ethanol peak after 3 min retention time, set against the HPLC chromatogram of pure gasoline (without ethanol). From Fig. 3, it is clear that there is no interference (superimposition) of the ethanol and gasoline peaks at this retention time to prevent the quantification of the anhydrous ethanol fuel content in the blend. The chromatogram of the gasoline matrix (filled line) does not peak at the same retention time as the ethanol (dotted line). In addition, methanol cannot be detected if methanol is the mobile phase, showing that the HPLC method proposed is selective for ethanol. The quantity of anhydrous ethanol fuel determined for this sample using the reference method (ABNT NBR 13992/2008) was 25 vol%, because both methanol and ethanol are water-soluble. However, the proposed HPLC method found the ethanol content to be 11.97 vol % ± 0.43 vol%. In other words, it was not affected by the presence of an interferent, which puts it at an advantage over the reference method currently used in Brazil. The expanded uncertainty of measurement of the HPLC method for determining the level of ethanol in the gasoline based on the standard samples used to plot the analytical curve was calculated at 0.56 vol% (Fig. 4S, supplementary material). The uncertainty of measurement in HPLC is a quantitative indication of the quality of the result obtained, and is therefore a tool that makes the analytical results and thus the analyses conducted using HPLC more reliable. In making the calculation, the contributions of the following sources of uncertainty were considered: the calibration of the pipettes and flasks used to prepare each standard sample; the purity of the ethanol used; the straight-line equation obtained; the preparation of the control sample; and the repeatability of the measurements. The spreadsheet showing the uncertainty of measurement calculations for this HPLC method is included in the Supplementary material (Fig. 4S).
1000
6
21
Fig. 2. Analytical curve showing ethanol concentration (% volume) against area (µRIU*min).
To plot the analytical curve, the six standard samples of gasoline containing anhydrous ethanol fuel (nominal values: 20 vol%, 22 vol%, 24 vol%, 26 vol%, 28 vol%, and 30 vol%) were analyzed by HPLC in quadruplicate, generating a dispersion graph in Excel of ethanol concentration (% volume) versus anhydrous ethanol fuel peak area (µRIU*min), as shown in Fig. 2. As the coefficient of determination (R2 = 0.9910) indicated a linear correlation, the equation of the analytical curve (y = 2.1608 x + 17.978) was used to quantify the ethanol concentrations determined by HPLC in the 12 verification samples for the partial validation of the proposed method. Table 2 shows the results of the HPLC analyses of the anhydrous ethanol fuel content in the 12 verification samples performed in quadruplicate. The results in Table 2 show that the method was valid with regard to accuracy, since all the relative errors obtained in the 12 verification samples were lower than 5%, the level set as the maximum admissible error (95% confidence interval). The results also show that the total amplitude of the measurements was lower than the repeatability limit in all the verification samples analyzed, which indicates the method was valid in terms of repeatability. Fig. 3 shows the HPLC chromatogram of the standard sample
4
70
ethanol on entra on (% volume)
3.2. Partial validation of proposed method
2
75
19
All the samples analyzed by HPLC had the same chromatographic profile, as shown in Fig. 1. The anhydrous ethanol fuel was eluted in around 3 min, unlike the constituent parts of the gasoline, whose retention times ranged from 4 min to 10 min.
0
80
60
3.1. Analysis of samples by high-performance liquid chromatography (HPLC)
-200
y = 2.1608x + 17.978 R² = 0.991
65
3. Results and discussion
1200
Analy cal Curve
85
16
Time (min)
Fig. 1. HPLC chromatogram of a standard sample containing 28% (vol) ethanol.
238
Fuel 212 (2018) 236–239
L.M. Avila et al.
Table 2 Anhydrous ethanol fuel content in the verification samples, determined by HPLC. Sample
Nominal Ethanol Value (% volume)*
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12
Anhydrous Ethanol Fuel Content (% volume)
19.01 19.99 20.99 22.98 22.95 24.99 25.96 26.96 26.95 31.04 32.06 39.99
Mean of Quadruplicate Measurements (% volume)
Standard Deviation
Absolute Error
Relative Error (%)
Repeatability Limit
Amplitude
19.76 19.51 21.84 23.57 23.33 25.45 26.38 27.28 27.61 30.08 30.88 38.55
0.04 0.20 0.96 0.85 0.09 0.18 0.34 0.50 0.11 0.50 0.31 1.17
0.75 0.48 0.84 0.60 0.37 0.46 0.42 0.32 0.66 0.96 1.18 1.44
3.95 2.41 4.02 2.61 1.63 1.84 1.63 1.20 2.46 3.08 3.68 3.60
0.12 0.55 2.70 2.38 0.25 0.50 0.95 1.41 0.30 1.41 0.88 3.27
0.10 0.42 2.01 1.75 0.22 0.43 0.71 1.18 0.22 1.11 0.76 2.69
* Corrected nominal values.
1000 900 12,5% EtOH + 12,5% MeOH
Height (μRIU)
800
[11]
700
Pure Gasoline
600
[12]
500 400
[13]
300 200
[14]
100 0 2.8
2.9
3.0
3.1
[15]
3.2
Time (min)
[16]
Fig. 3. HPLC chromatograms of a standard sample containing 12.5 vol% ethanol and 12.5 vol% methanol and a sample of pure gasoline.
[17]
Acknowledgements
[18]
The authors wish to thank Companhia Brasileira de Petróleo Ipiranga and ANP for the gasoline supplied.
[19]
[20]
Appendix A. Supplementary data
[21]
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.fuel.2017.10.039.
[22]
References
[23]
[1] [2] [3] [4]
[5] [6] [7]
[8] [9]
[10]
Portaria MAPA N°75, March 5, 2015. Brazil, 2015. http://www.anp.gov.br, (accessed in February 2017). ANP Resolution N°40, October 25, 2013. Brazil, 2013. NBR 13992/2015. Gasolina Automotiva - Determinação do Teor de Etanol Anidro Combustível [Automotive Gasoline - Determination of Anhydrous Ethanol Fuel Content]. 2015. http://economia.estadao.com.br/noticias/geral,anp-descobre-nova-adulteracao-doalcool-hidratado,4518e, (accessed in February 2017). http://oglobo.globo.com/economia/petroleo-e-energia/anp-encontra-metanol-emalcool-hidratado-para-veiculos-20467071, (accessed in February 2017). http://blogs.oglobo.globo.com/ancelmo/post/anp-encontra-16-milhoes-de-litrosde-combustivel-adulterado-em-tanques-de-shell-ipiranga-e-br.html, (accessed in February 2017). http://odia.ig.com.br/rio-de-janeiro/2016-11-16/anp-interdita-postos-da-ilha-quevendiam-alcool-adulterado.html, (accessed in February 2017). E.C.S. Oliveira, A. Cunha Neto, R.C. Silva, E.V.R. Castro, V. Lacerda Jr., W. Romão, R.G. Pereira, T. Sten, Determinação de Adulteração de Etanol Combustível com Metanol por Ressonância Magnética Nuclear Associada a Ferramentas Quimiométricas [Determination of Adulteration of Ethanol Fuel with Methanol by Nuclear Magnetic Resonance Associated with Chemometric Tools], in: 52nd Brazilian Congress of Chemistry, Recife, Brazil, 2012. Available at: http://www. abq.org.br/cbq/2012/trabalhos/9/845-14325.html. Guilherme BCM, Mateus AM, Paulo AZS. Kit Colorimétrico para Detecção de
[24] [25]
[26]
[27]
[28]
[29] [30]
239
Metanol em Etanol Combustível para o Monitoramento da Qualidade de Combustíveis [Colorimetric Kit for Detection of Methanol in Ethanol Fuel for Monitoring Fuel Quality]. Quim Nova 2015;38:280–4. http://dx.doi.org/10.5935/ 0100-4042.20140305. ASTM D5501/2016. Standard Test Method for Determination of Ethanol and Methanol Content in Fuels Containing Greater than 20% Ethanol by Gas Chromatography. 2012. (Reapproved in 2016). NBR 16041. Etanol Combustível - Determinação dos Teores de Metanol e Etanol por Cromatografia Gasosa [Ethanol Fuel - Determination of Methanol and Ethanol Content by Gas Chromatography]. 2015. Corsetti S, Zehentbauer FM, McGloin D, Kiefer J. Characterization of gasoline/ ethanol blends by infrared and excess infrared spectroscopy. Fuel 2015;141:136–42. Fernandes HL, Raimundo Jr. IM, Pasquini C, Rohwedder JJR. Simultaneous determination of methanol and ethanol in gasoline using NIR spectroscopy: effect of gasoline composition. Talanta 2008;75:804–10. Conklin Jr. A, Goldcamp MJ, Barrett J. Determination of ethanol in gasoline by FTIR spectroscopy. J Chem Educ 2014;91:889–91. Fortunato FM, Vieira AL, Gomes Neto JA, Donati GL, Jones BT. Expanding the potentialities of standard dilution analysis: determination of the ethanol in gasoline by Raman spectroscopy. Microchem J 2017;133:76–80. Kumar K, Mishra AK. Quantification of ethanol in ethanol-petrol and biodiesel in biodiesel-diesel blends using fluorescence spectroscopy and multivariate methods. J Fluoresc 2012;22:339–47. Pereira PF, Marra MC, Munoz RAA, Richter EM. Fast batch injection analysis system for on-site determination of ethanol in gasohol and fuel ethanol. Talanta 2012;90:99–102. Arik E, Altan H, Esenturk O. Dielectric properties of ethanol and gasoline mixtures by terahertz spectroscopy and an effective method for determination of ethanol content of gasoline. J Phys Chem 2014;118:3081–9. Turanov A, Khitrin AK. Proton NMR characterization of gasoline-ethanol blends. Fuel 2014;137:335–8. Pereira PF, Souza RMF, Munoz RAA, Richter EM. Simultaneous determination of ethanol and methanol in fuel ethanol using cyclic voltammetry. Fuel 2013;103:725–9. http://dx.doi.org/10.1016/j.fuel.2012.07.034. Abreu REL, Paz JEM, Silva AC, Pontes MJC, Lemos SG. Ethanol fuel adulteration with methanol assessed by cyclic voltammetry and multivariate calibration. Fuel 2015;156:20–5. Lutz OMD, Bonn GK, Rode BM, Hucka CW. Reproducible quantification of ethanol in gasoline via a customized mobile near-infrared spectrometer. Anal Chim Acta 2014;826:61–8. ASTM D5599/2010. Test Method for Determination of Oxygenates in Gasoline by Gas Chromatography and Oxygen Selective Flame Ionization Detection. 2010. ASTM D4815/2013. Standard Test Method for Determination of MTBE, ETBE, TAME, DIPE, Tertiary-Amyl Alcohol and C1 to C4 Alcohols in Gasoline by Gas Chromatography. 2013. Lopes PA. Validação de Métodos de Ensaios para Laboratórios de Ensaios de Calibração [Validation of Test Methods for Calibration Laboratories]: Internal publication of method validation course. Rio de Janeiro, Brazil: Rede de Tecnologia; 2008. Rego ECP. Validação de Métodos de Ensaios Químicos [Validation of Chemical Testing Methods]: Internal publication of method validation course. Rio de Janeiro, Brazil: Instituto Nacional de Metrologia, Qualidade e Tecnologia (INMETRO); 2013. Ribeiro FAL, Ferreira MMC, Morano SC, Silva LR, Schneider RP. Planilha de Validação: Uma Nova Ferramenta para Estimar Figuras de Mérito na Validação de Métodos Analíticos Univariados [Validation Spreadsheet: a new tool for estimating figures of merit for the validation of univariate methods.]. Química Nova 2008;31:164–71. http://dx.doi.org/10.1590/S0100-40422008000100029. http://www.inmetro.gov.br/Sidoq/Arquivos/Cgcre/DOQ/DOQ-Cgcre-8_04.pdf, (accessed in February 2017). http://www.inmetro.gov.br/inovacao/publicacoes/gum_final.pdf, (accessed in March 2015).
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Full Length Article
Determination of ethanol in gasoline by high-performance liquid chromatography
MARK
⁎
Lorena Morine Avilaa, Amanda Pereira Franco dos Santosb, , Danielle Ignácio Mançano de Mattosb, Cristiane Gimenes de Souzab, Débora França de Andradec, Luiz Antonio d'Avilab a b c
Instituto de Química, Universidade Federal Fluminense, Niterói, Rio de Janeiro 24020-141, Rio de Janeiro, Brazil Engineering Program in Chemical and Biochemical Processes, Escola de Química, Universidade Federal do Rio de Janeiro, 21941-909 Rio de Janeiro, Brazil Instituto de Química, Universidade Federal do Rio de Janeiro, Cidade Universitária, 21941-909 Rio de Janeiro, Brazil
A R T I C L E I N F O
A B S T R A C T
Keywords: Gasoline Ethanol HPLC
This study employed non-aqueous reversed-phase high-performance liquid chromatography (HPLC) with refractive-index detection and methanol as a mobile phase to quantify the ethanol content in gasoline. The advantages of HPLC are its good separation versatility, high resolution, relatively short analysis time, and automation. Standard samples were prepared to obtain a standard curve and partial validation. Accuracy, precision, linearity, selectivity for methanol (adulterant), and measurement uncertainty were investigated. In the partial validation analysis, the accuracy of the proposed method was identified as its main advantage over the reference method.
1. Introduction When fuel is adulterated, it is usually done by adding a lower cost product in order to obtain illicit financial gains from its retail. The illegal addition of excess ethanol to gasoline is arguably the easiest and most common form of adulteration, since ethanol is already a component of the gasoline/ethanol blend used in Brazil (27% ± 1% by volume) and is cheaper than gasoline. According to the national bulletin of the Fuel Quality Monitoring Program run by the Brazilian fuel regulatory agency (ANP), the ethanol content of gasoline is the biggest cause of noncompliance found in samples of gasoline [1,2]. The reference method for determining ethanol content, described in ANP resolution N°40 of October 25, 2013, is detailed in Brazilian standard NBR 13992/2015, issued by the Brazilian technical standards association (Associação Brasileira de Normas Técnicas, ABNT) [3,4]. It is a quick, simple, practical method that can be done in the field to check gasoline purity. However, because of these very features, its measurement uncertainty is 1 vol%, and it has the added limitation of quantifying other water-soluble alcohols like methanol, impairing its accuracy. The adulteration of ethanol by methanol has been verified in Brazil by the ANP [5–8], and this has motivated the development of tests to determine methanol content in ethanol [9,10].
The American Society for Testing and Materials (ASTM) has its own standard, D5501, published in 2012 and reapproved in 2016, which describes a standard method using gas chromatography (GC) for the determination of the ethanol (20 mass% or over) and methanol (0.01 mass%–0.6 mass%) content of fuels [3,11]. This method is not so easy to implement, as it is done using long columns (100 m and 150 m), which, while they do exist, are harder to find on the market and very expensive. In the case of the 100 m column, the initial temperature of the column is supposed to be 15 °C, but this is unfeasible in much of Brazil as it is lower than ambient temperature, meaning that a cooling stage would first have to be employed. Further, the specific density of all the samples must be calculated to correct the results obtained [11]. In Brazil, ABNT set up a working group at its Ethanol Fuel Study Commission to create a reference method for quantifying ethanol and methanol in gasoline and ethanol fuel. The chosen method was GC, but after much research no reference method for quantifying ethanol in gasoline was produced, since the compounds in gasoline are eluted in the same time as ethanol [12]. There are different methods described in the literature for analyzing oxygenates in gasoline for quality control and regulatory purposes, like Fourier-transform near-infrared spectroscopy [13] with partial least square (PLS) calibration [14,15]; Raman spectroscopy [16]; synchronous fluorescence spectroscopy with principal component regression or
⁎
Corresponding author at: Centro de Tecnologia, Bloco K02, LABCOM, Cidade Universitária, Rio de Janeiro 21941-909, Brazil. E-mail addresses: [email protected] (L.M. Avila), [email protected] (A.P.F. dos Santos), [email protected] (D.I.M. de Mattos), [email protected] (C.G. de Souza), [email protected] (D.F. de Andrade), [email protected] (L.A. d'Avila). http://dx.doi.org/10.1016/j.fuel.2017.10.039 Received 21 February 2017; Received in revised form 22 September 2017; Accepted 6 October 2017 0016-2361/ © 2017 Elsevier Ltd. All rights reserved.
Fuel 212 (2018) 236–239
L.M. Avila et al.
Calibrated glassware (10.00 mL volumetric flask, 1.00 mL volumetric pipette, and 10–100 µL variable volume automatic micropipette) was also used in the preparation of this sample.
PLS calibration models [17]; a combination of excitation-emission matrix fluorescence spectroscopy with multiway partial least square regression (N-PLS) and unfolded PLS [17]; batch injection analysis with detection by multiple-pulse amperometry [18]; Terahertz spectroscopy [19]; proton nuclear magnetic resonance spectroscopy (1H NMR) [20]; cyclic voltammetry and multivariate calibration [21,22]; and customized mobile near-infrared spectrometry [23]. ASTM D5599/2010 describes a procedure for determining oxygenate content in gasoline by GC, but only for the 0.1–20% by mass range, which is not enough for the Brazilian fuel market [24]. ASTM D4815/2013 describes a GC method for determining methyl tert-butyl ether (MTBE), ethyl tert-butyl ether (ETBE), diisopropyl ether (DIPE), tert-amyl alcohol, and C1–C4 alcohols in gasoline. However, this method is employed to determine alcohol levels of 0.2 mass% to 12 mass%, ruling out its use for ethanol-based fuels and making it unfeasible for testing Brazilian gasolines [25]. In this context, this study proposes the use of high-performance liquid chromatography (HPLC) with refractive index detection to develop a quick method for quantifying the ethanol content in gasoline. It is hoped that this method may replace the existing reference test for laboratory analyses and the monitoring and inspection of gasoline sold in Brazil, whose drawbacks are its minimum uncertainty of ± 1% (by volume) and the fact that it is not accurate enough for determining ethanol levels in gasoline if it is adulterated with methanol.
2.2. Analysis of the samples by high-performance liquid chromatography (HPLC) All the samples were analyzed using non-aqueous reversed-phase high-performance liquid chromatography (HPLC) with a Dionex UltiMate™ 3000 quaternary pump (Thermo Scientific, Massachusetts, USA), a Shodex RI-101 refractive index detector (ECOM, Czech Republic), and an UltiMate™ WPS-3000 autosampler (Thermo Scientific) with a 100 µL sample loop. An Acclaim™ column (Thermo Scientific) measuring 250 mm long and 4.6 mm internal diameter was used with an octadecylsilane phase with 5 µm particle size and 120 Å pore diameter. The mobile phase was 100% pre-filtered HPLC-grade methanol (Tedia Brazil, Rio de Janeiro, Brazil). Total analysis time was 15 min at 40 °C (temperature of column oven), with a constant flow rate of 1 mL/min and 10 µL injection volume. The samples were injected in quadruplicate to evaluate repeatability. The chromatograms were analyzed using Chromeleon 6.80 SR11 software (Thermo Scientific) with manual integration. 2.3. Partial validation of proposed method In order to use objective evidence to ascertain whether the proposed method produces reliable results that are fit for purpose – i.e. whether it meets the requirements and standards for the specific proposed use – we assessed its linearity, selectivity for methanol (adulterant), accuracy, precision, and measurement uncertainty [26–30]. Linearity was checked by the coefficient of determination (R2) of the straight line. Values over 0.99 indicated the linear working range, where the angular coefficient of the straight line could be considered constant [26]. The selectivity of the HPLC method for methanol was checked by injecting a sample of pure gasoline (matrix) and a standard sample containing 12.5 vol% ethanol, 12.5 vol% methanol, and 75 vol% gasoline, since no certified reference material exists for such a determination. In order to assess the accuracy of the verification samples, we used Eq. (1) to ascertain relative error (Erel) [26]:
2. Materials and methods 2.1. Preparation of standard samples of gasoline with ethanol Eighteen standard samples of gasoline containing different quantities of anhydrous ethanol fuel, ranging from 19 to 40% (by volume), were prepared. The gasoline was supplied by Companhia Brasileira de Petróleo Ipiranga and the ANP inspection department, and support was provided by the Fuel and Petroleum Products Laboratory at the School of Chemistry, Federal University of Rio de Janeiro (LABCOM/EQ/ UFRJ). All the glassware used in preparing the samples (1.00, 2.00, 5.00, and 10.00 mL volumetric pipettes; 10–100 µL variable volume automatic micropipette; and 50.00 mL volumetric flask) were calibrated in advance. Six of the 18 samples were used to plot an analytical curve and 12 were used to verify the analytical curve and to partially validate the proposed method, as shown in Table 1. Aside from the 18 standard samples, a sample of gasoline containing 12.5 vol% ethanol and 12.5 vol% methanol was prepared to verify accuracy when ethanol was adulterated with methanol and to demonstrate its selectivity for ethanol.
Erel =
Standard Samples for Verification
Sample Code
Nominal Ethanol Value (% volume)*
Sample Code
Nominal Ethanol Value (% volume)*
A1 A2 A3 A4 A5 A6
20.00 22.00 24.00 26.00 28.00 30.00
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12
19.00 20.00 21.00 23.00 23.00 25.00 26.00 27.00 27.00 31.00 32.00 40.00
(1)
where: Erel is relative error (%); Xlab is the single value obtained experimentally or the mean of the laboratory results; Xv is the value accepted as being true.
Table 1 Standard samples of gasoline containing different proportions of anhydrous ethanol fuel. Standard Samples for Analytical Curve
xlab−x v ∗100 xv
In this study, the relative error of the 12 verification samples (V1–V12) was calculated, assuming a 95% confidence interval for the statistical treatment of the results obtained. As such, the maximum admissible error was set at 5%. The repeatability limit (r) calculated by Eq. (2) and the total amplitude of the measurements (difference between the highest and lowest value from a dataset) were used to evaluate the precision of the proposed method [27].
r = 2.8∗Sr
(2)
where: r is the repeatability limit; Sr is the standard deviation of repeatability, meaning the standard deviation of the results for each concentration.
* The nominal values were corrected using the data from the glassware calibration certificates (see Table 2, in Results).
237
Fuel 212 (2018) 236–239
L.M. Avila et al.
Finally, the expanded uncertainty of measurement of the HPLC method described here, for a 95% confidence interval, was calculated according to the Guide to the Expression of Uncertainty in Measurement (ISO GUM) [30]. It considered the following sources of uncertainty: the calibration of the pipettes and flasks used to prepare each standard sample; the purity of the ethanol used; the straight-line equation obtained; the preparation of the control sample; and the repeatability of the measurements. For each source of uncertainty, we ascertained what type of uncertainty was involved and what method was best suited to determining and calculating it. We carried out type A and type B evaluations of uncertainty and closely observed the contributions made by more than one source of uncertainty so that no calculations were duplicated. Electronic spreadsheets were used for the calculations of combined uncertainty, effective number of degrees of freedom, coverage factor, and expanded uncertainty. The spreadsheet showing the uncertainty of measurement calculations for this HPLC method is included in the Supplementary material.
Height (μRIU)
area (μRIU*min)
600 400 200 0 12
14
27
29
31
The proposed high-performance liquid chromatography method proved effective for quantifying the anhydrous ethanol fuel content of gasoline for its speed, simplicity, accuracy, and precision, demonstrated by the statistical tools used in its partial validation. However, its greatest advantage is the accuracy it was found to have, as it suffered no interference from the other constituent parts of the matrix (gasoline) or the interferent (methanol). This is one of the drawbacks of the traditional method and the main reason why this research was done. We conclude that the HPLC method described in this study is promising for the accurate determination of the ethanol content of gasoline. The expanded uncertainty of this method (0.56 vol%) is also lower than that of the current reference test (1.00 vol%).
800
10
25
4. Conclusion
Ethanol
8
23
containing 12.5 vol% ethanol and 12.5 vol% methanol in gasoline A, highlighting the ethanol peak after 3 min retention time, set against the HPLC chromatogram of pure gasoline (without ethanol). From Fig. 3, it is clear that there is no interference (superimposition) of the ethanol and gasoline peaks at this retention time to prevent the quantification of the anhydrous ethanol fuel content in the blend. The chromatogram of the gasoline matrix (filled line) does not peak at the same retention time as the ethanol (dotted line). In addition, methanol cannot be detected if methanol is the mobile phase, showing that the HPLC method proposed is selective for ethanol. The quantity of anhydrous ethanol fuel determined for this sample using the reference method (ABNT NBR 13992/2008) was 25 vol%, because both methanol and ethanol are water-soluble. However, the proposed HPLC method found the ethanol content to be 11.97 vol % ± 0.43 vol%. In other words, it was not affected by the presence of an interferent, which puts it at an advantage over the reference method currently used in Brazil. The expanded uncertainty of measurement of the HPLC method for determining the level of ethanol in the gasoline based on the standard samples used to plot the analytical curve was calculated at 0.56 vol% (Fig. 4S, supplementary material). The uncertainty of measurement in HPLC is a quantitative indication of the quality of the result obtained, and is therefore a tool that makes the analytical results and thus the analyses conducted using HPLC more reliable. In making the calculation, the contributions of the following sources of uncertainty were considered: the calibration of the pipettes and flasks used to prepare each standard sample; the purity of the ethanol used; the straight-line equation obtained; the preparation of the control sample; and the repeatability of the measurements. The spreadsheet showing the uncertainty of measurement calculations for this HPLC method is included in the Supplementary material (Fig. 4S).
1000
6
21
Fig. 2. Analytical curve showing ethanol concentration (% volume) against area (µRIU*min).
To plot the analytical curve, the six standard samples of gasoline containing anhydrous ethanol fuel (nominal values: 20 vol%, 22 vol%, 24 vol%, 26 vol%, 28 vol%, and 30 vol%) were analyzed by HPLC in quadruplicate, generating a dispersion graph in Excel of ethanol concentration (% volume) versus anhydrous ethanol fuel peak area (µRIU*min), as shown in Fig. 2. As the coefficient of determination (R2 = 0.9910) indicated a linear correlation, the equation of the analytical curve (y = 2.1608 x + 17.978) was used to quantify the ethanol concentrations determined by HPLC in the 12 verification samples for the partial validation of the proposed method. Table 2 shows the results of the HPLC analyses of the anhydrous ethanol fuel content in the 12 verification samples performed in quadruplicate. The results in Table 2 show that the method was valid with regard to accuracy, since all the relative errors obtained in the 12 verification samples were lower than 5%, the level set as the maximum admissible error (95% confidence interval). The results also show that the total amplitude of the measurements was lower than the repeatability limit in all the verification samples analyzed, which indicates the method was valid in terms of repeatability. Fig. 3 shows the HPLC chromatogram of the standard sample
4
70
ethanol on entra on (% volume)
3.2. Partial validation of proposed method
2
75
19
All the samples analyzed by HPLC had the same chromatographic profile, as shown in Fig. 1. The anhydrous ethanol fuel was eluted in around 3 min, unlike the constituent parts of the gasoline, whose retention times ranged from 4 min to 10 min.
0
80
60
3.1. Analysis of samples by high-performance liquid chromatography (HPLC)
-200
y = 2.1608x + 17.978 R² = 0.991
65
3. Results and discussion
1200
Analy cal Curve
85
16
Time (min)
Fig. 1. HPLC chromatogram of a standard sample containing 28% (vol) ethanol.
238
Fuel 212 (2018) 236–239
L.M. Avila et al.
Table 2 Anhydrous ethanol fuel content in the verification samples, determined by HPLC. Sample
Nominal Ethanol Value (% volume)*
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12
Anhydrous Ethanol Fuel Content (% volume)
19.01 19.99 20.99 22.98 22.95 24.99 25.96 26.96 26.95 31.04 32.06 39.99
Mean of Quadruplicate Measurements (% volume)
Standard Deviation
Absolute Error
Relative Error (%)
Repeatability Limit
Amplitude
19.76 19.51 21.84 23.57 23.33 25.45 26.38 27.28 27.61 30.08 30.88 38.55
0.04 0.20 0.96 0.85 0.09 0.18 0.34 0.50 0.11 0.50 0.31 1.17
0.75 0.48 0.84 0.60 0.37 0.46 0.42 0.32 0.66 0.96 1.18 1.44
3.95 2.41 4.02 2.61 1.63 1.84 1.63 1.20 2.46 3.08 3.68 3.60
0.12 0.55 2.70 2.38 0.25 0.50 0.95 1.41 0.30 1.41 0.88 3.27
0.10 0.42 2.01 1.75 0.22 0.43 0.71 1.18 0.22 1.11 0.76 2.69
* Corrected nominal values.
1000 900 12,5% EtOH + 12,5% MeOH
Height (μRIU)
800
[11]
700
Pure Gasoline
600
[12]
500 400
[13]
300 200
[14]
100 0 2.8
2.9
3.0
3.1
[15]
3.2
Time (min)
[16]
Fig. 3. HPLC chromatograms of a standard sample containing 12.5 vol% ethanol and 12.5 vol% methanol and a sample of pure gasoline.
[17]
Acknowledgements
[18]
The authors wish to thank Companhia Brasileira de Petróleo Ipiranga and ANP for the gasoline supplied.
[19]
[20]
Appendix A. Supplementary data
[21]
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.fuel.2017.10.039.
[22]
References
[23]
[1] [2] [3] [4]
[5] [6] [7]
[8] [9]
[10]
Portaria MAPA N°75, March 5, 2015. Brazil, 2015. http://www.anp.gov.br, (accessed in February 2017). ANP Resolution N°40, October 25, 2013. Brazil, 2013. NBR 13992/2015. Gasolina Automotiva - Determinação do Teor de Etanol Anidro Combustível [Automotive Gasoline - Determination of Anhydrous Ethanol Fuel Content]. 2015. http://economia.estadao.com.br/noticias/geral,anp-descobre-nova-adulteracao-doalcool-hidratado,4518e, (accessed in February 2017). http://oglobo.globo.com/economia/petroleo-e-energia/anp-encontra-metanol-emalcool-hidratado-para-veiculos-20467071, (accessed in February 2017). http://blogs.oglobo.globo.com/ancelmo/post/anp-encontra-16-milhoes-de-litrosde-combustivel-adulterado-em-tanques-de-shell-ipiranga-e-br.html, (accessed in February 2017). http://odia.ig.com.br/rio-de-janeiro/2016-11-16/anp-interdita-postos-da-ilha-quevendiam-alcool-adulterado.html, (accessed in February 2017). E.C.S. Oliveira, A. Cunha Neto, R.C. Silva, E.V.R. Castro, V. Lacerda Jr., W. Romão, R.G. Pereira, T. Sten, Determinação de Adulteração de Etanol Combustível com Metanol por Ressonância Magnética Nuclear Associada a Ferramentas Quimiométricas [Determination of Adulteration of Ethanol Fuel with Methanol by Nuclear Magnetic Resonance Associated with Chemometric Tools], in: 52nd Brazilian Congress of Chemistry, Recife, Brazil, 2012. Available at: http://www. abq.org.br/cbq/2012/trabalhos/9/845-14325.html. Guilherme BCM, Mateus AM, Paulo AZS. Kit Colorimétrico para Detecção de
[24] [25]
[26]
[27]
[28]
[29] [30]
239
Metanol em Etanol Combustível para o Monitoramento da Qualidade de Combustíveis [Colorimetric Kit for Detection of Methanol in Ethanol Fuel for Monitoring Fuel Quality]. Quim Nova 2015;38:280–4. http://dx.doi.org/10.5935/ 0100-4042.20140305. ASTM D5501/2016. Standard Test Method for Determination of Ethanol and Methanol Content in Fuels Containing Greater than 20% Ethanol by Gas Chromatography. 2012. (Reapproved in 2016). NBR 16041. Etanol Combustível - Determinação dos Teores de Metanol e Etanol por Cromatografia Gasosa [Ethanol Fuel - Determination of Methanol and Ethanol Content by Gas Chromatography]. 2015. Corsetti S, Zehentbauer FM, McGloin D, Kiefer J. Characterization of gasoline/ ethanol blends by infrared and excess infrared spectroscopy. Fuel 2015;141:136–42. Fernandes HL, Raimundo Jr. IM, Pasquini C, Rohwedder JJR. Simultaneous determination of methanol and ethanol in gasoline using NIR spectroscopy: effect of gasoline composition. Talanta 2008;75:804–10. Conklin Jr. A, Goldcamp MJ, Barrett J. Determination of ethanol in gasoline by FTIR spectroscopy. J Chem Educ 2014;91:889–91. Fortunato FM, Vieira AL, Gomes Neto JA, Donati GL, Jones BT. Expanding the potentialities of standard dilution analysis: determination of the ethanol in gasoline by Raman spectroscopy. Microchem J 2017;133:76–80. Kumar K, Mishra AK. Quantification of ethanol in ethanol-petrol and biodiesel in biodiesel-diesel blends using fluorescence spectroscopy and multivariate methods. J Fluoresc 2012;22:339–47. Pereira PF, Marra MC, Munoz RAA, Richter EM. Fast batch injection analysis system for on-site determination of ethanol in gasohol and fuel ethanol. Talanta 2012;90:99–102. Arik E, Altan H, Esenturk O. Dielectric properties of ethanol and gasoline mixtures by terahertz spectroscopy and an effective method for determination of ethanol content of gasoline. J Phys Chem 2014;118:3081–9. Turanov A, Khitrin AK. Proton NMR characterization of gasoline-ethanol blends. Fuel 2014;137:335–8. Pereira PF, Souza RMF, Munoz RAA, Richter EM. Simultaneous determination of ethanol and methanol in fuel ethanol using cyclic voltammetry. Fuel 2013;103:725–9. http://dx.doi.org/10.1016/j.fuel.2012.07.034. Abreu REL, Paz JEM, Silva AC, Pontes MJC, Lemos SG. Ethanol fuel adulteration with methanol assessed by cyclic voltammetry and multivariate calibration. Fuel 2015;156:20–5. Lutz OMD, Bonn GK, Rode BM, Hucka CW. Reproducible quantification of ethanol in gasoline via a customized mobile near-infrared spectrometer. Anal Chim Acta 2014;826:61–8. ASTM D5599/2010. Test Method for Determination of Oxygenates in Gasoline by Gas Chromatography and Oxygen Selective Flame Ionization Detection. 2010. ASTM D4815/2013. Standard Test Method for Determination of MTBE, ETBE, TAME, DIPE, Tertiary-Amyl Alcohol and C1 to C4 Alcohols in Gasoline by Gas Chromatography. 2013. Lopes PA. Validação de Métodos de Ensaios para Laboratórios de Ensaios de Calibração [Validation of Test Methods for Calibration Laboratories]: Internal publication of method validation course. Rio de Janeiro, Brazil: Rede de Tecnologia; 2008. Rego ECP. Validação de Métodos de Ensaios Químicos [Validation of Chemical Testing Methods]: Internal publication of method validation course. Rio de Janeiro, Brazil: Instituto Nacional de Metrologia, Qualidade e Tecnologia (INMETRO); 2013. Ribeiro FAL, Ferreira MMC, Morano SC, Silva LR, Schneider RP. Planilha de Validação: Uma Nova Ferramenta para Estimar Figuras de Mérito na Validação de Métodos Analíticos Univariados [Validation Spreadsheet: a new tool for estimating figures of merit for the validation of univariate methods.]. Química Nova 2008;31:164–71. http://dx.doi.org/10.1590/S0100-40422008000100029. http://www.inmetro.gov.br/Sidoq/Arquivos/Cgcre/DOQ/DOQ-Cgcre-8_04.pdf, (accessed in February 2017). http://www.inmetro.gov.br/inovacao/publicacoes/gum_final.pdf, (accessed in March 2015).