Electronic Supplementary Material
The BIOREMA project – part 2: International interlaboratory comparison for biodiesel test methods
Manuela Ulberth-Buchgraber(*)1, Monica Potalivo 1,2, Andrea Held1, Annarita Baldan3, Adriaan M.H. van der Veen3, Hugo Ent3, Valnei S. Cunha 4, Romeu J. Daroda4, Brian Lang 5, Michele Schantz5, Ruth Hearn 6, Richard J. C. Brown 7, Paul J. Brewer 7
1European Commission, Joint Research Centre, Institute for Reference Materials and Measurements (IRMM), Retieseweg 111, 2440 Geel, Belgium
2Current address: Istituto Superiore per la Protezione e la Ricerca Ambientale (ISPRA), Servizio Metrologia Ambientale, Via di Castel Romano 100, 00128 Roma, Italy
3VSL, Thijsseweg 11, 2629JA Delft, The Netherlands
4 Chemical Metrology Division, National Institute of Metrology, Quality and Technology (INMETRO), Av. Nossa Senhora das Graças 50, CEP 25250020.RJ. Brazil
5Biochemical Sciences Division and Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, MD20899USA
6 LGC, Queens Road, Teddington, TW11 0LY, UK
7 Analytical Science Division, National Physical Laboratory, Hampton Road, Teddington, TW11 0LW, UK
(*) Corresponding author: Manuela Ulberth-Buchgraber
Tel+32 14571819
Fax+32 14571548
E-mail:
Table 1: Consensus value and statistical parameters of the results of the interlaboratory comparison: n: number of datasets after removal of outliers, sr: relative repeatability standard deviation; sL relative between-laboratory standard deviation; sR: relative reproducibility standard deviation
Analyte / n / Outliers removed / Consensus value / sr [%] / sL [%] / sR [%]Density at 15 °C [kg/m3] / 23 / 4 / 883.467 / 0.003 / 0.01 / 0.01
Viscosity at 40 °C [mm2/s] / 21 / 3 / 4.4492 / 0.18 / 0.33 / 0.38
Water content [mg/kg] / 19 / 1 / 216 / 5.09 / 10.19 / 11.57
Sulfur content [mg/kg] / 15 / 0 / 2.16 / 10.19 / 18.06 / 20.83
Methanol content [10-2 g/g] / 16 / 1 / 0.0216 / 11.57 / 18.98 / 22.22
Total ester content [10-2 g/g] / 18 / 0 / 97.61 / 0.39 / 1.19 / 1.25
Linolenic acid ME content [10-2 g/g] / 16 / 3 / 9.9 / 0.91 / 1.41 / 1.72
C16:0 content [10-2 g/g] / 16 / 1 / 4.64 / 1.51 / 2.16 / 2.59
C18:0 content [10-2 g/g] / 16 / 1 / 1.7 / 2.35 / 5.29 / 5.88
Total sum of C18:1
content [10-2 g/g] / 16 / 4 / 59.06 / 0.41 / 0.54 / 0.68
Total sum of C18:2
content [10-2 g/g] / 16 / 4 / 20.38 / 1.03 / 0 / 1.03
Total sum of C18:3
content [10-2 g/g] / 15 / 3 / 9.86 / 1.01 / 2.33 / 2.54
Iodine value expressed as mass of iodine used per 100 g of sample [g] / 18 / 2 / 113.5 / 0.97 / 1.85 / 2.11
Flash point [°C] / 17 / 0 / 175.7 / 1.02 / 5.07 / 5.18
Oxidation stability [h] / 17 / 2 / 12 / 0.83 / 3.33 / 4.17
Acid value expressed as mass KOH used per 1 g of sample [g] / 19 / 2 / 0.188 / 9.57 / 18.62 / 20.74
Table 2: Results,consensus value and statistical parameters of the results obtained in the interlaboratory comparison for glycerides and glycerol:n: number of datasets after removal of outliers, sr: relative repeatability standard deviation; sL relative between-laboratory standard deviation; sR: relative reproducibility standard deviation; n.p.: data not provided by laboratory
Analyte / Monoacylglycerols [10-2 g/g] / Diacyl-glycerols [10-2 g/g] / Triacyl-glycerols [10-2 g/g] / Total glycerol [10-2 g/g] / Monolein [10-2 g/g] / Diolein [10-2 g/g] / Triolein [10-2 g/g]Lab code
L001 / 0.67 / 0.18 / 0.12 / 0.21 / n.p. / n.p. / n.p.
L002 / 0.67 / 0.15 / 0.06 / 0.2 / 0.56 / 0.07 / 0.06
L003 / 0.64 / 0.12 / 0.05 / 0.19 / n.p. / n.p. / n.p.
L004 / 0.65 / 0.13 / 0.03 / 0.19 / 0.55 / 0.1 / 0.02
L005 / 0.72 / 0.1 / 0.05 / 0.21 / 0.66 / 0.05 / n.p.
L006 / n.p. / n.p. / n.p. / n.p. / n.p. / n.p. / n.p.
L007 / 0.65 / 0.08 / 0.02 / 0.18 / n.p. / n.p. / n.p.
L008 / n.p. / n.p. / n.p. / n.p. / n.p. / n.p. / n.p.
L009 / n.p. / n.p. / n.p. / n.p. / n.p. / n.p. / n.p.
L010 / 0.66 / 0.15 / 0.07 / 0.2 / n.p. / n.p. / n.p.
L011 / 0.71 / 0.09 / 0.05 / 0.2 / 0.61 / 0.04 / 0.05
L012 / 1.03 / 0.17 / 0.05 / 0.3 / n.p. / n.p. / n.p.
L013 / 0.76 / 0.16 / 0.03 / 0.23 / 0.69 / 0.12 / 0.03
L014 / 0.53 / 0.18 / 0.02 / 0.16 / n.p. / n.p. / n.p.
L015 / 0.03 / 0.01 / n.p. / 404.83 / 0.03 / 0 / n.p.
L016 / n.p. / n.p. / n.p. / n.p. / n.p. / n.p. / n.p.
L017 / n.p. / n.p. / n.p. / n.p. / n.p. / n.p. / n.p.
L018 / 0.56 / 0.13 / 0.06 / 0.17 / 0.52 / 0.11 / 0.06
L019 / 0.96 / 0.16 / 0.07 / 0.28 / 0.88 / 0.1 / 0.04
L020 / 0.64 / 0.08 / n.p. / 0.18 / n.p. / n.p. / n.p.
L021 / 0.71 / 0.09 / 0.04 / 0.2 / n.p. / n.p. / n.p.
L022 / 0.6 / 0.16 / 0 / 0.18 / 0.6 / 0.13 / 0
L023 / 0.87 / 0.14 / 0.07 / 0.25 / n.p. / n.p. / n.p.
Consensus value / 0.66 / 0.13 / 0.05 / 0.2 / 0.63 / 0.08 / 0.03
No. of outliers removed / 3 / 1 / 0 / 2 / 1 / 0 / 0
sr [%] / 2.58 / 4.48 / 20 / 2.99 / 3.63 / 3.75 / 14.71
sl [%] / 10.45 / 25.37 / 60 / 14.43 / 17.85 / 53.75 / 61.76
sR [%] / 10.76 / 25.37 / 60 / 14.93 / 18.17 / 53.75 / 61.76
Kinematic viscosity, which impacts fuel injection and combustion, is closely related tothe fatty acid composition of a given biodiesel sample. It increases with increasing length of both the fatty acid chain and the alcohol group and is inversely related to the number of double bonds[1].Kinematic viscosity is useful in monitoring the fuel quality of biodiesel during storage since it continuously increases with decreasing fuel quality [2]. The reference value (given in Table 1 in the main article) for the BIOREMA test material B is based on the weighted mean of the results provided by three NMIs using different methods. In the interlaboratory comparison, 21 laboratories reported results for kinematic viscosity, from which three were excluded from the calculation of the consensus value(L012, L015, and L016). From the remaining datasets 10fall within the 95 % coverage interval of the reference value (Figure 1). In principle, ASTM D6751 [3]requires the use of ASTM D445[4], and EN 14214 [5] utilises ISO 3104 [6]. No trend was visible for the individual methodsused in the interlaboratory comparison. Tenlaboratories usedEN ISO 3104[6], sixlaboratoriesapplied ASTM D445[4], and one reported a combination of thesetwo methods. Furthermore, one laboratory measured according to ASTM D7042[7], using a Stabinger viscosimeter, whichshould provide results equivalent to ISO 3104[6]/ASTM D445[4]. Another laboratory followed the standard method IP71[8],which is supposedly technically equivalent to the ASTM D445[4], but the laboratory results were flagged asdiscrepant. Twolaboratories did not state their measurement method. Overall, there isa good agreement between the reference and the consensus values, with a relative difference of less than 0.02 % and no significant influences of the individual methods could be observed.
Figure 1.Laboratory results for viscosity. The uncertainty bar represents two times the standard deviation of four measurement results. The solid line indicates the reference value with its expanded uncertainty (dashed lines).
Watercan be introduced into biodieselduring the final washing step of the production process and has to be reduced by drying. Even very low water contents achieved directly after production do not guarantee however that biodiesel fuels will still meet the specifications during combustion. As biodiesel is hygroscopic, it can absorb water during handling and storage.Once the solubility limit is exceeded, water separates inside the storage tank and collects at the bottom [1]. Water in biodiesel can promote microbial growth, tank corrosion, formation of emulsions, and cause hydrolysis or hydrolytic oxidation[2]. In the BIOREMA project, the reference value for water content was assigned by using three independent datasets from different NMIs, two of them using a coulometric Karl Fischer titration and one using a volumetric Karl Fischer titration. 19 laboratories participated in the interlaboratory comparison using various methods, i.e.,12 applied ISO 12937[9], 4used ASTM D6304 [10]and one used ASTM D1744[11], which has been withdrawn. Anotherlaboratorystated IP438[12], which should be equivalent to ISO 12937[9]. The results from the laboratory using IP438 [12]were not in line with the remaining results however and consequently excluded from the calculation of the consensus value. One laboratory did not report the measurement method.Figure 2shows that there is no meaningful influence of the methods. The agreement between the consensus and reference valuesis perceived as satisfactory with a relative deviation of less than 1 %. Moreover, 17 laboratories report results within the 95 % coverage interval around the assigned value. Unfortunately, the expanded standard uncertainty of the assigned value was ofthe same order as the reproducibility standard deviation stated in the standard test method ISO 12937[9], due to a higher standard uncertainty from the between-bottle homogeneity. Further research with respect to material processing is needed in order to attain a smaller uncertainty for the assigned value.
Figure 2. Laboratory results for water content. The uncertainty bar represents two times the standard deviation of four measurement results. The solid line indicates the reference value with its expanded uncertainty (dashed lines).
Methanol is a raw material used to produce FAME. Methanol remaining in the product, such as from incomplete reactions during production or contamination, can have adverse effects, such as lowered flash point, decreased lubricity, corroded injectors and degraded materials in fuel distribution and vehicle fuel systems, all of them raising handling and safety concerns. According to EN 14214[5], the methanol content must not exceed 0.2·10-2 g/g in biodiesel, and the recommended standard analytical procedure (EN 14110[13]) is based on capillary gas chromatography in combination with flame ionisation detection using a polar stationary phase [1]. In the characterisation study, two NMIs determined the methanol content. The data failed a consistency check using aχ2-test [14] with a 99 % probability. As the results with their uncertainties were not overlapping, the arithmetic mean was calculated for the value assignment, which did not qualify as reference value. Moreover, the evaluated standard uncertainty due to between-bottle homogeneity was substantially largerthan the reproducibility standard deviation stated in the documentary standard EN 14110[13],thus indicating issues in the material processing step.However, it was difficult to judge whetherthe issues arise from the processing of the material, the measurements carried out, or both.Of the 16 laboratories participating in the intercomparison study, 12used EN 14110[13], one applied NBR 15343[15], also a GC-based method, and three did not report their measurement method. One laboratory using EN 14110 [13]was found discrepant (L016)and excluded from the consensus value calculation. The consensus value, based on 15 measurement results,was lower than the mean from the NMI results, but they still agreed within their respective uncertainties, mainly owing to thelarge uncertainty associated with the mean value of the NMI results.
Figure3.Laboratory results for methanol content. The uncertainty bar represents two times the standard deviation of four measurement results. The solid line indicates the reference value with its expanded uncertainty (dashed lines).
The iodine value is a measure of total unsaturation of a lipidic material. The determination according to EN 14111[16] in the European biodiesel standard is based on the classic wet chemical method, a titrimetric method using the Wijs reagent (iodine monochloride dissolved in acetic acid). In the BIOREMA project,the value assignment of the iodine value was carried out by 11expert laboratories using EN 14111[16], i.e., the measurand was method-defined. Of the 18 laboratories participating in the comparison, 16 used EN 14111[16], and two did not state their method. Two datasets (L012, L015) were found discrepant and excluded from calculating the consensus value. In both cases the use of a certified reference material would have been very helpful as quality control. Finally, there isa good agreement between the reference and the consensus values (Figure 4). Furthermore, 11laboratoriesare within the 95 % coverage interval around the reference value.
Figure 4. Laboratory results for iodine value. The uncertainty bar represents two times the standard deviation of four measurement results. The solid line indicates the reference value with its expanded uncertainty (dashed lines).
Oxidation stability is one of the most important properties because biodiesel fuels are more sensitive to oxidative degradation than fossil diesel fuel. This higher sensitivity is especially true for fuels with a high content of polyunsaturated esters, as the methylene groups adjacent to double bonds are particularly susceptible to oxidation [1]. The reference value in the project, established by 10 expert laboratories using EN 14112[17], is method-defined.Of the 17 laboratories participating in the interlaboratory comparison, 15 used the same method (EN 14112[17]), one used EN 15751[18], which is also based on EN 14112[17], and one did not state its measurement method. Two datasets were found discrepant (L007, L022) and eliminated from the calculation of the consensus value. The distribution of participants' results is narrow, with a between laboratory standard deviation (sL) of less than 3.5 % relative. Moreover, the agreement between the reference and the consensus valuesis very good and 13 laboratories are within the 95 % coverage interval around the reference value (Figure 5).
Figure 5. Laboratory results for oxidation stability. The uncertainty bar represents two times the standard deviation of four measurement results. The solid line indicates the reference value with its expanded uncertainty (dashed lines).
References
[1]Mittelbach M. Remschmidt C. (2006) Biodiesel: the comprehensive handbook. 3rd Edn. M. Mittelbach,Graz, Austria
[2]Knothe G (2006) Analyzing biodiesel: Standards and other methods. JAOCS 83: 823-833
[3]ASTM D6751 (2011) Standard Specification for Biodiesel Fuel Blend Stock (B100) for Middle Distillate Fuels. American Society for Testing and Materials, West Conshohocken
[4]ASTM D445 (2009) Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity). American Society for Testing and Materials, West Conshohocken
[5]EN 14214 (2008+A1:2009) Automotive fuels: Fatty acid methyl esters (FAME) for diesel engines: Requirements and test methods. European Committee for Standardization, Brussels
[6]ISO 3104 (1994) Petroleum products - Transparent and opaque liquids - Determination of kinematic viscosity and calculation of dynamic viscosity. International Organization for Standardization, Geneva
[7]ASTM D7042 (2004) Standard Test Method for Dynamic Viscosity and Density of Liquids by Stabinger Viscometer (and the Calculation of Kinematic Viscosity). American Society for Testing and Materials, West Conshohocken
[8]IP 71 Section 1 (1997) Petroleum products -Transparent and opaque liquids - Determination of kinematic viscosity and calculation of dynamic viscosity. (Technically equivalent standard: ASTM D 445-03). Institute of Petroleum Standards, GB
[9]ISO 12937 (2000) Petroleum products - Determination of water - Coulometric Karl Fischer titration method. International Organization for Standardization, Geneva
[10]ASTM D6304 (2007) Standard Test Method for Determination of Water in Petroleum Products, Lubricating Oils, and Additives by Coulometric Karl Fischer Titration. American Society for Testing and Materials, West Conshohocken
[11]ASTM D1744 (1992) Standard Test Method for Determination of Water in Liquid Petroleum Products by Karl Fischer Reagent (Withdrawn 2000). American Society for Testing and Materials, West Conshohocken
[12]IP 438 (2001) Petroleum products - Determination of water - Coulometric Karl Fischer titration method (Identical Standards: BS 2000: Part 438: 2001; EN ISO 12937: 2000; ISO 12937: 2000). Institute of Petroleum Standards, GB
[13]EN 14110 (2003) Fat and oil derivatives. Fatty acid methyl esters (FAME). Determination of methanol content. European Committee for Standardization, Brussels
[14]Cox MG (2002) The evaluation of key comparison data. Metrologia 39: 589-595
[15]NBR 15343 (2008) Biodiesel – Determination of methanol or/and ethanol concentrations in fatty acid esters (biodiesels) by gas chromatography. Brazilian Association for Technical Standards, Rio de Janeiro
[16]EN 14111 (2003) Oil and fat derivatives - Fatty Acid Methyl Esters (FAME) - Determination of iodine value. European Committee for Standardization, Brussels
[17]EN 14112 (2003) Fat and oil derivatives. Fatty acid methyl esters (FAME). Determination of oxidation stability (accelerated oxidation test). European Committee for Standardization, Brussels
[18]EN 15751 (2009) Automotive fuels - Fatty acid methyl ester (FAME) fuel and blends with diesel fuel - Determination of oxidation stability by accelerated oxidation method. European Committee for Standardization, Brussels
1
1