A Rapid and Quantitative HPLC Method for Determination of Diethylene Glycol

Abstract

A method was developed for determining diethylene glycol (DEG) by HPLC. Good separation was achieved by analysis of standard mixtures containing normal wine components and DEG. Addition of DEG to wine samples proved that DEG did not elute with any of the known constituents. Recovery studies proved the method to have a good repeatability. A selection of South African and imported wines were analysed and it was found that none contained any DEG. Since the start of the Austrian wine "scandal" (Frei-hold, 1985; Krieghauser, 1985; Lord, 1985) it has become important to be able to determine the presence of diethylene glycol (DEG) in wines. Freihold (1985) and Pfeiffer and Radler (1985) recently gave account of the properties, toxicity and need to determine DEG. Published methods for the detection and identification of DEG include the use of packed column gas chromatography (GC) (Wagner & Kreuzer, 1985), capillary column GC and GC-mass spectrometry (Bandion, Valen-ta & Kohlmann, 1985; Haase-Aschoff & Haase-Aschoff, 1985) and thin-layer chromatography (Anon., 1985; Lehmann & Ganz, 1985). Bertrand (1985) used a complicated, time consuming extraction technique and-capillary column GC to quantify DEG determination in wme. We found the slightly modified HPLC method reported by Schwarzenbach (1982), which we used for quantitative determination of organic acids in wines, to be suitable for the detection of small amounts of DEG. A similar finding was reported by Pfeiffer & Radler (1985) on completion of this study. This paper concerns the quantitative HPLC determination of DEG and the DEG-status of a number of South African and imported wines. MATERIALS AND METHODS Wine samples: One hundred wine samples were obtained from the Sub-Directorate: Quality Control of which 22 were imported from Germany and Italy, ranging from the vintages 1977 to 1983. The remaining 78 wines were submitted to the South African Wine and Spirit Board for certification. of which 18 were intended for export and the remaining 60 for local distribution. All imported wines contained less than 30 gil residual sugar. Wines intended for export included 7 dry white (less than 4 gil residual sugar), 5 semi-sweet and late harvest (more than 4 but less than 30 gil residual sugar), 6 special late harvest and noble late harvest (more than 20 but less than 50 gil residual sugar for special late harvest and more than 50 gil residual sugar for noble late harvest) wines. The wines intended for local distribution included 26 dry. 29 semi-sweet and late harvest. 4 special late harvest and noble late harvest and 2 dessert wines (sweet, fortified). Standards: Standard solutions containing 20, 15, 10, 5, and 1 gil of diethylene glycol (Merck 803131) in freshly distilled and deionized water were prepared. A mixture of 20 gil each of glucose and fructose (Merck 8342 and 5323), 9 gil glycerol (May & Baker 73969), 0,5 gil acetic acid (Riedel De Haen AG 33209), 1 gil2,3 buthylene glycol (Fluka 18970), diethylene glycol (Merck 822329) and 10% (viv) ethanol (redistilled, NCP) was also prepared. For the recovery study of diethylene glycol (DEG) from different wine types 20, 15, 10,5 and 1 gil of DEG was added to respectively a semi-sweet. a late harvest. a special late harvest and a sweet fortified dessert wine. To determine the sensitivity of the method, three wine types viz. a special late harvest, a late harvest and a semi-sweet wine were spiked with respectively 1 gil, 0,1 gil and 0,()1 gil of DEG. A 100 ml aliquot was concentrated fourfold by evaporation on a boiling water bath. Chromatography: Except for the auto sampler (Spark Holland), a Knauer High Performance Liquid Chromatography system (HPLC) was used. This included a Model 64.00 pump and a Model 98.00 differential refractometer. Integration was done with an Apple Ile fitted with Chro-matochart (Interactive Microware, Inc. P A 16801) chromatography software. The column used was an Aminex HPX-87H 300 x 7,8 mm ion exclusion column (Bio Rad 1250140) fitted with a Microguard ion exclusion column 40 x 4,6 mm (Bio Rad 1250129). Operating Conditions: Mobile phase, freshly distilled water was deionized to 17 Megohms at 20°C by means of a Nanopure (Barn-stead) and acidified to 0,013 N with H,PO, (Merck 565); column temperature, 50°C (constant); flow rate. 1,0 ml/min; injection volume. 20 11-l. differential refrac-tometer, x8; run time. 16 min. Sample Preparation: Five ml aliquots of wine and standard samples were passed through a Waters C18-Sep Pak clean up preco-lumn (Subden et al., 1979). followed by a 0,2 11-m mem-The authors express their sincere appreciation to Mrs M. Jespers for technical assistance. S. Afr.