Universal detectors for determination of lipids in biodiesel production

By Marc Plante, Deanna C. Hurum, Ian Acworth, and Jeffrey S. Rohrer

In This Section

October 2010

In recent years, biodiesel has been promoted as a renewable and environmentally preferable fuel option to petroleum diesel. Biofuels, including biodiesel, are predicted to replace over 10 billion gallons (38 billion liters) of the petroleum currently used in the United States by 2030. Biodiesel is often produced by reacting plant or animal oils with an alcohol to form the desired alkyl esters as fuel, and this transesterification is typically achieved using a base. This also produces free glycerol and acylglycerols as impurities in the biodiesel that must be limited in the final fuel product because they form harmful deposits in engines and damage emissions control systems.

Standard Requirements

To ensure reliable biodiesel quality, ASTM International has adopted ASTM D6751, which is largely equivalent to the European Union standard, EN 14214. D6751 applies to B100 biodiesel (i.e., fuel that is 100% biodiesel, with no petrodiesel) that is used for blending, ensuring that the source material and biodiesel blends are of high quality. Among the many parameters controlled by this standard are the residual concentrations of glycerol and acylglycerols in the product.

Many methods to determine glycerols in biodiesel samples exist. The currently recommended method is a high-temperature gas chromatography (GC) method, ASTM D6584. However, this GC method requires sample derivatization and the use of high-temperature columns, which degrade quickly. With the goal of simplifying sample preparation, HPLC (high-performance liquid chromatography) has been investigated for both acylglycerol and total glycerol determination in biofuels.

HPLC Solutions

Because acylglycerols are a broad class of compounds, gradient HPLC methods are preferred both to improve separation and shorten analytical run times. However, many of the solvents used to separate these analytes, such as tetrahydrofuran and ethyl acetate, absorb in the UV (ultraviolet) spectrum, causing large changes in the UV baseline, which makes UV detection of analytes challenging. Additionally, glycerols lack a strong chromophore, inherently limiting method sensitivity. For these reasons, other HPLC detection schemes have been investigated.

ELSD and CAD

Nebulizer-based universal detectors are becoming more popular for the determination of acylglycerols in lipids and biofuels. These detectors do not rely on analyte chromophores. Instead, detection is based on analyte mass. For these detectors, the HPLC eluent is nebulized, creating an aerosol. Once the mobile phase is evaporated, solid analyte particles remain, from which the detector can determine the amount of analyte that is present. The two methods of detection that are commonly used are evaporative light scattering detection (ELSD) and charged aerosol detection (CAD). These detectors are limited to nonvolatile and some semivolatile compounds.

ELSD uses light scattering to determine analyte mass and therefore concentration in the original sample. Analyte response is nonlinear resulting from the contribution of different light-scattering mechanisms producing a sigmoidal response curve. ELSD instruments generally provide moderate sensitivity and precision.

In comparison, CAD charges the particles for analyte determination as depicted in Figure 1. Following mobile phase evaporation, the analyte particles collect charge from a stream of charged nitrogen molecules, formed when nitrogen gas passes over a charged corona wire. The charge on the particles is measured by a sensitive electrometer, which provides the signal (current). CAD typically shows similar inter-analyte response, uniform response curves, and greater sensitivity, precision, accuracy, and wider dynamic range (>4 orders) than ELSD.

Examples of Analytical Separations

Many methods can be used to analyze lipids in biodiesel feedstocks and final products. The samples can be characterized with high resolution and with rapid analysis times, depending on the analytical needs. For example, Figure 2 illustrates the high-resolution separation of algal oil, as a representative fuel feedstock, using HPLC-CAD. This chromatogram clearly shows the high sensitivity of the detector for the minor lipids contained within the sample. This method can separate and quantify a wide variety of different lipid classes, from steroids to triacylglycerols.

Figure 3A shows a rapid (10 min) HPLC-ELSD separation of a B99 sample produced from recycled restaurant fryer oil. In this chromatogram, the fatty acid methyl esters elute first, followed by diacylglycerols, and finally triacylglycerols. The sensitivity of ELSD, with LOQ (limits of quantification) of 8-15 µg/mL, does allow for some acylglycerol determinations. However, it is difficult to determine acylglycerols in this sample under these conditions. In contrast, when the same sample is analyzed by Corona ultra, as shown in Figure 3B, the increased sensitivity easily allows glycerol oleate quantification. With CAD, the estimated LOQ for mono-, di-, and trioleoylglycerols are 1.5-2.4 µg/mL, showing ample sensitivity to determine acylglycerols below the EN limits. Additionally, for comparison, the Corona ultra peak area precision is dramatically improved compared to ELSD with relative standard deviations of 0.2-0.6% for 50 µg/mL standards compared to 2.3-4.8%, respectively. Tables 1 and 2 detail the improved calibration, LOQ, and precision values for the model acylglycerols with Corona ultra detection compared to ELSD.

An analyte that is of considerable importance in biodiesel analysis, which is not quantified with the above methods, is free glycerol. To determine free and bound acylglycerols in biodiesel, another HPLC-CAD method is available that will provide quantification of total glycerols in a single run. Although not presented here, readers are referred to Application Note #70-8305 on www.coronaultra.com  for specific information.

Conclusions

Determining lipids is a challenge in any matrix. Low UV absorbance and low volatility of larger lipids limit the sensitivity of existing HPLC and GC methods. Universal HPLC detectors are showing promise for the determination of these analytes, as compounds need not possess a chromophore, and they do not require derivatization. By using HPLC-CAD, both biofuel feedstocks and products can be analyzed quickly, with good precision, and with the necessary sensitivity to meet existing standards. CAD is the preferred detector owing to its sensitivity, wide dynamic range, and similar interanalyte response.

Marc Plante, Deanna C. Hurum, Ian Acworth, and Jeffrey S. Rohrer are all with Dionex Corp. (Sunnyvale, California, USA). Contact Marc Plante via email at marc.plante@dionex.com.

information

For further reading:

Plante, M., B. Bailey, and I. Acworth, The use of charged aerosol detection with HPLC for the measurement of lipids, in Lipidomics: Vol. 1: Methods and Protocols, Vol. 579 of series Methods in Molecular Biology, edited by D. Armstrong, Humana Press, New Jersey, 2009, pp. 469-482.