Rapid determination of acrylamide contaminant in conventional fried foods by GC

Rapid determination of acrylamide contaminant in conventional fried foods by GC, Rapid determination of acrylamide contaminant in conventional fried foods by GC, Rapid determination of acrylamide contaminant in conventional fried foods by GC, Rapid determination of acrylamide contaminant in conventional fried foods by GC

Journal of Chromatography A, 1116 (2006) 209–216 Rapid determination of acrylamide contaminant in conventional fried foods by gas chromatography with electron capture detector Yu Zhang a , Yi Dong a , Yiping Ren b , Ying Zhang a,∗ a Department of Food Science and Nutrition, College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310029, Zhejiang Province, China b Zhejiang Provincial Center for Disease Prevention and Control, Hangzhou 310009, Zhejiang Province, China Received December 2005; received in revised form March 2006; accepted March 2006 Available online April 2006 Abstract Gas chromatography coupled with electron capture detector (GC–ECD) was successfully developed and applied for the rapid determination of acrylamide in conventional fried foods, such as potato crisps, potato chips, and fried chicken wings The method included defatting with n-hexane, extraction with aqueous solution of sodium chloride (NaCl), derivatization with potassium bromate (KBrO3 ) and potassium bromide (KBr), and liquid–liquid extraction with ethyl acetate The final acrylamide extract was analyzed by GC–ECD for quantification and by GC–MS for confirmation The chromatographic analysis was performed on the HP-INNOWax capillary column, and good retention and peak response of acrylamide were achieved under the optimal conditions (numbers of theoretical plates N = 83,815) The limit of detection (LOD) was estimated to be 0.1 ␮g kg−1 on the basis of ECD technique Recoveries of acrylamide from conventional samples spiked at levels of 150, 500 and 1000 ␮g kg−1 (n = for each level) ranged between 87 and 97% with relative standard deviations (RSD) of less than 4% Furthermore, the GC–ECD method showed that no clean-up steps of acrylamide derivative would be performed prior to injection and was slightly more sensitive than the MS/MS-based methods Validation and quantification results demonstrated that this method should be regarded as a new, low-cost, and robust alternative for conventional investigation of acrylamide © 2006 Elsevier B.V All rights reserved Keywords: Acrylamide; Conventional fried foods; Electron capture detection; Derivatization Introduction Acrylamide, a well-known neurotoxic compound, was detected in carbohydrate-rich fried or baked food samples by the research group from Swedish National Food Administration (SNFA) and University of Stockholm in 2002 [1] General findings suggest that acrylamide be likely found in foods that have been processed by various heat-treated methods other than boiling One possible pathway to the formation of acrylamide in some foods is possibly via the Maillard reaction, which involves the reaction of a specific amino acid with a reducing sugar in the presence of heat [2,3] It has been shown that the likely reactants which produce significant levels of acrylamide in foods are asparagine and reducing sugar However, 2-deoxyglucose is a weak source for the formation of acrylamide because 2- ∗ Corresponding author Tel.: +86 571 8697 1388; fax: +86 571 8604 9803 E-mail address: y zhang@zju.edu.cn (Y Zhang) 0021-9673/$ – see front matter © 2006 Elsevier B.V All rights reserved doi:10.1016/j.chroma.2006.03.042 deoxyglucose does not have a hydroxyl group adjacent to the carbonyl, and it can only form Schiff base adduct and cannot undergo the Amadori rearrangement, which leads to the formation of dicarbonyl compounds, e.g 3-deoxyglucosone [4] Another possible route involves the reaction of asparagine with octanal to form acrylamide These findings suggest the necessity of carbonyls in the formation of acrylamide, but that of dicarbonyls are not essential [4,5] In addition, a report by Yasuhara et al [6] suggested that in lipid-rich foods the reaction of acrylic acid from acrolein and ammonia from the amino acid upon heating may also play an important role in acrylamide formation Although considerable controversy exists regarding the exposure levels relevant to carcinogenicity of acrylamide in humans, the reports on the presence of acrylamide in European food prompted the U.S Food and Drug Administration (FDA) to analyze a variety of foods sold in the United States for the presence of acrylamide Recently in 2005, World Health Organization (WHO) and Food and Agriculture Organization (FAO) together announced that certain foods processed 210 Y Zhang et al / J Chromatogr A 1116 (2006) 209–216 or cooked at high temperature, especially conventional snacks, contain considerable levels of acrylamide and may cause concern regarding risks to human health [7] Under such situation, researches on the analytical method and risk assessment of acrylamide in different food matrixes have once again become a hotspot A great number of hitherto published methods have been developed in the past years to quantitatively analyze the acrylamide monomer, including different extraction and clean-up procedures for different food matrixes Rosen and Hellenăas [8] firstly reported the analysis of acrylamide in different heat-treated foods using the isotope dilution liquid chromatography–tandem mass spectrometry (LC–MS/MS) technique They developed a mass spectrometry method for direct detection of acrylamide, which would unequivocally verify the presence of acrylamide in a range of heat-treated foods Tareke et al [9] systematically made the investigations of acrylamide in heated foodstuffs by comparison between GC–MS and LC–MS/MS Besides these contributions, a sequence of analytical methods dealing with the analysis of acrylamide in heat-treated foods has been published in peer-review journals, reported by specific research groups or presented at international scientific conferences [10] The gas chromatograph (GC) step is based on derivatization (usually bromination to increase volatility) of the analyte or on direct analysis without derivatization As for the method sensitivity, limits of quantification (LOQ) range from 30 to 50 ␮g kg−1 for LC–MS down to 10 to 30 ␮g kg−1 for GC–MS For the detection of acrylamide levels below 30 ␮g kg−1 , GC–MS with bromination is to be favoured over the non-derivatized method [11] As for further optimization, Roach et al [12] described an analytical method employing LC–MS/MS that reduced the limit of quantification down to 10 ␮g kg−1 with a recovery rate of 95% An updated analytical LC–MS/MS method for ground and brewed coffee was published [13] with limits of detection (LOD) of 10 ␮g kg−1 and ng ml−1 , respectively Hitherto, it can be summarized from recent methodological studies that GC–MS and LC–MS/MS appeared to be acknowledged as the most useful and popular method for acrylamide determination [11,14–19] Although MS is chosen as a main technique for GC-based analysis, the mass of acrylamide derivatives or their fragment ions is not specific owing to the presence of co-extractives that yield the same magnitude of m/z with acrylamide derivatives in the sample matrix These interferences could not be completely avoided in these methods despite the use of MS owing to poor retention of polar acrylamide molecule on conventional GC capillary columns Therefore, the efforts from researchers focused on purifying the extracts and removing the co-extractives via clean-up optimization Most clean-up procedures consisted of the combination of several solid-phase extractions (SPE) via various cartridges [5,20,21] Such clean-up procedures cost too much time to improve the pretreatment efficiency When a huge number of food samples need to be investigated, the general pretreatment steps including impurity removal, analyte extraction, derivatization, and clean-up prior to MS analysis by previous published studies could not satisfy the requirements of rapid determination of acrylamide Under such situation, the establishment of a reliable, sensitive, rapid, and low-cost analytical method for the determination of acrylamide becomes a prerequisite Recently, a sensitive reverse-phase HPLC-DAD method for acrylamide analysis in potato-based processed foods was developed by Găokmen et al [22] and the rapid and convenient measurement was successfully achieved Similarly, a robust GC method based on conventional detection other than MS should be developed and optimized in order to fulfill the target of rapid analysis The objective of the present study is to develop and validate a rapid, highly sensitive, and reliable GC method for the determination of acrylamide in conventional fried foods The method is developed on the basis of derivatization of the target analyte with bromination and detection by electron capture detector (ECD), which can be easily adopted by nonspecialized analytical laboratories The sample pretreatment steps include the extraction of acrylamide from sample matrixes by sodium chloride solution, defatting process by n-hexane, and derivatization with potassium bromate (KBrO3 ) and potassium bromide (KBr) without clean-up prior to GC analysis The results obtained from GC–ECD analysis are confirmed by GC–MS Experimental 2.1 Chemicals Acrylamide (99%) and 13 C3 -labeled acrylamide (isotopic purity 99%) were purchased from Sigma–Aldrich (St Louis, MO, USA) and Cambridge Isotope Laboratories (Andover, MA, USA), respectively All of other solvents and chemicals such as potassium bromide, potassium bromate, and sodium chloride used for the analysis of acrylamide were of analytical grade Ethyl acetate and n-hexane should be redistilled before use Water was purified with a Milli-Q system (Millipore, Bedford, USA) Acrylamide and 13 C3 -labeled acrylamide are potent cumulative neurotoxins in animals and men and may be carcinogenic They are hazardous and should be handled carefully Sample pretreatment procedures referring to organic reagent operations should be carried out in a fume cupboard 2.2 Samples Several representative samples of conventional fried foods were purchased from two major Western food suppliers in Hangzhou, China The analytical survey comprised a sequence of commercial products such as potato crisps, potato chips, fried chicken wings, etc Considering protecting commercial benefits of Western food suppliers, brands of tested samples were not shown and pointed out Certified reference test material (FAPAS T3011 potato crisps) was obtained from Central Science Laboratory (Sand Hutton, York, UK) to validate the robustness of the present method Y Zhang et al / J Chromatogr A 1116 (2006) 209–216 2.3 Sample preparation The fried food samples were pulverized and homogenized in HL-2070 multi-function food processor (Shanghai Herine Electric Appliance Co Ltd., Shanghai, China) or WH861 variable speed Waring blender (Taicang Science and Educational Instrument Factory, Taicang, Jiangsu, China) prior to sampling As for sampling, 1.50 g of the above-mentioned samples was weighed in 50 ml centrifuge tubes For GC–MS confirmatory tests, the sample matrixes were additionally spiked with 500 ␮l of 13 C3 labeled acrylamide (1 ␮g ml−1 ) as the internal standard and the tubes were placed for 10 in order that labeled acrylamide could adequately mix with sample matrix via osmotic effect To make a defatting process, 20 ml of redistilled n-hexane was added, and each tube was then capped and shaken by hand or vortex briefly to mix the contents of tube The tubes were then clamped and shaken in KQ3200E ultrasonic shaker (Kunshan Ultrasonic Instrument Co Ltd., Kunshan, Jiangsu, China) to mix the tube contents for 10 The supernatant n-hexane was removed, and the defatting step was then performed again as described above Seven milliliters of sodium chloride (2 mol l−1 ) were added into the residue of each tube, which was capped and shaken in an ultrasonic shaker to extract the analyte for 20 The tubes were centrifuged at 15,000 rpm for 15 with a Microfuge 18 Beckman Coulter centrifuge (Beckman Coulter Inc., Fullerton, CA, USA) The clarified aqueous layer was promptly removed by a pipet The residues were extracted again by ml of sodium chloride, and the extraction step was performed as described above The supernatant fluids during two extraction steps were removed by a pipet and merged for further use Five milliliters of the subsamples of the aqueous solution from NaCl extraction and 0.6-ml sulfuric acid (10%, v/v) were sequentially added into a brown quantitative colorimetric cylinder The volume of solution was quantitatively fixed up to 10 ml with the addition of NaCl solution and the cylinder was then placed into refrigerating cabinet for precooling (4 ◦ C, 15 min) An aliquot of derivatization reactants, including ml of 0.1-mol l−1 potassium bromate and 1.50 g of potassium bromide powder, was added to the precooled solution The cylinder was shaken with a vortex blender, and the reaction mixture was allowed to stand for 30 at ◦ C The derivatization reaction was terminated by adding 0.1 ml of 0.1-mol l−1 sodium thiosulfate A 4-ml aliquot of analyte solution was extracted thrice with ml of redistilled ethyl acetate, and the combined extracts were dried over sodium sulfate One microliter of the final test solution was injected onto GC capillary column for both quantification by GC–ECD or confirmation by GC–MS 211 automatic liquid sampler and injector system (Agilent) onto a 19091N-113 HP-INNOWax capillary column (polyethylene glycol, 30-m length, 0.32-mm i.d., 25-␮m film thickness, J&W Scientific, Agilent, CA, USA) Separations were performed using nitrogen as the carrier gas, applying the following temperature program: 110 ◦ C (hold time min), then at 10 ◦ C min−1 to 140 ◦ C (hold time 15 min), and at 30 ◦ C min−1 to the final temperature of 240 ◦ C (hold time min) The GC–ECD sample injector interface temperature and detector interface temperature were both held at 250 ◦ C 2.5 Confirmatory analysis by GC–MS Brominated sample extracts prepared by aforementioned pretreatment steps and terminated by adding sodium thiosulfate were submitted to an additional SPE clean-up step so that the derivatized extracts could be purified, and co-extractive interference could obviously be reduced prior to MS-based analysis SPE clean-up was performed via OASIS HLB cartridges (200 mg, cm3 ) purchased from Waters Technology (Milford, MA, USA) The cartridges were conditioned with 3.5 ml of methanol followed by 3.5 ml of water; the methanol and water portions were discarded Each cartridge was loaded with 1.5 ml of brominated extract The extract was allowed to pass through the sorbent material and discarded Then, the cartridge was eluted with ml of water, and the eluent was collected and extracted thrice with ml of redistilled ethyl acetate The combined extracts were dried over sodium sulfate Finally, ␮l of the final test solution was injected onto GC–MS, which was performed on a Hewlett-Packard (HP) 6890 gas chromatograph coupled with an HP 5973 benchtop mass selective detector (MSD) operated in selected ion monitoring (SIM) mode with positive electron impact (EI) ionization The analytical separation was performed on a HP5-MS capillary column, (polysiloxane polymers, 30 m × Ø0.25 mm, 0.25 ␮m, J&W Scientific, Agilent, CA, USA) and helium was chosen as the carrier gas at a flow rate of 1.0 ml min−1 Following injection, the column was held at 60 ◦ C for min, then programmed at 10 ◦ C min−1 to 200 ◦ C, and held for at 200 ◦ C (total run time: 21 min) Injections by the autosampler were made in splitless mode with a purge activation time of 1.0 and an injection temperature of 280 ◦ C The GC–MS interface transfer line was held at 280 ◦ C Under such conditions, the retention time of acrylamide and 13 C -acrylamide derivatives was 6.6 Ions monitored were m/z 70, 149, and 151 for 2-bromopropenamide, and m/z 110 and 154 for 2-bromo(13 C3 )-propenamide Results and discussion 2.4 Quantitative analysis by GC–ECD 3.1 Sample preparation and extraction GC–ECD was used for both method validation and quantification of acrylamide in the tested samples Pretreated samples were analyzed on the chromatographic system including a 6890N network gas chromatograph from Agilent Technologies (Palo Alto, CA, USA) coupled with Agilent’s 6890N microelectron capture detector connected on line Sample volume of ␮l (solvent ethyl acetate) was injected on-column with a 7683 Spiking recovery tests might not adequately reflect the matrix interactions of naturally embedded analytes It is necessary to estimate and compare extraction efficiencies from complex sample matrixes with different extraction solvents Therefore, extraction of acrylamide from spiked conventional food samples was evaluated for sample homogenates with purified water 212 Y Zhang et al / J Chromatogr A 1116 (2006) 209–216 and mol l−1 of sodium chloride (NaCl) Recoveries for individual acrylamide level quantified by GC–ECD in various sample matrixes using water and NaCl were (29–43%) and (68–90%), respectively The greatly improved extraction efficiency using NaCl aqueous solution is probably owing to partial denaturation of proteins and inhibition of emulsification process [15,21,23,24] 3.2 Modified derivatization of acrylamide GC separation demands derivatization of acrylamide, which is well done in most laboratories with hydrobromic acid (HBr) and saturated Br2 solution [5,10,14,25–27] The excess bromine is then removed by addition of sodium thiosulfate until the solution becomes colorless so that the derivative reaction is terminated As an alternative technique, derivatization with KBrO3 and KBr was applied in the present study according to Nemoto et al [28] with some modifications, on the basis of the following reaction: Fig GC–ECD chromatogram of acrylamide derivatives: (A) 2-bromopropenamide (2-BPA); (B) 2,3-dibromopropionamide (2,3-DBPA) Column: HP-INNOWax; carrier gas: nitrogen an already cleaned extract as a final step before instrumental analysis 3.3 Extraction of acrylamide derivatives using optimal volume of ethyl acetate KBrO3 + KBr + H2 SO4 → Br + K2 SO4 + H2 O CH2 CH–CO–NH2 + Br2 → CH2 Br–CHBr–CO–NH2 + CH2 CBr–CO–NH2 There are many advantages of this derivatization technique Use of the strong acid (HBr) and saturated Br2 solution can be avoided as these two solvents have been prepared difficultly and handled hazardously Nevertheless, use of KBrO3 and KBr combination is relatively more convenient and safe, and the reaction is performed in about 30 at cold storage temperature with excellent reproducibility, i.e less than 10% GC–ECD area difference of derivatized standard in three duplicates at three different days with the same concentration In addition, no further cleanup after derivatization is needed Meanwhile (KBrO3 + KBr) derivatives of acrylamide show excellent GC properties, i.e sharp peak shapes and good response in ECD As described above, the two derivatives, i.e 2,3-dibromopropionamide (2,3DBPA, 95%) are less polar compared with the original compound and are therefore, easily soluble in non-polar organic solvents like ethyl acetate The GC–ECD chromatogram of two derivatives is shown in Fig 1, and 2-BPA was chosen as the quantitative analyte because the peak response of 2-BPA was nearly 20 times higher than 2,3-DBPA Such phenomenon is partly due to the instability of 2,3-DBPA, which can be converted to the more stable derivative 2-BPA on the inlet of the GC or directly on the capillary column This decomposition (dehydrobromination) may yield poor repeatability and accuracy so that it is preferable to deliberately convert 2,3-DBPA to the stable 2-BPA prior to GC analysis, which can be readily done by adding 10% of triethylamine to the final extract before injection This conversion is almost instantaneous at room temperature and has been shown to be quantitative and reproducible [29,30] Disadvantages of (KBrO3 + KBr) derivatization are the complex composition of the solvent that could influence the reaction yield of the derivatives Such derivatization should therefore, be performed with Similar to the experiences of private and official food control laboratories, problems have been encountered in the analysis of difficult matrixes owing to interfering compounds in the characteristic acrylamide transitions for the analyte and difficultly in concentration Furthermore, the analyte should be extracted and transferred into the organic solvent phase prior to GC injection A promising approach is to extract the analyte into a polar organic solvent, such as ethyl acetate [4] The ethyl acetate extract could then be concentrated and injected On the basis of the above tests, it is necessary to demonstrate whether acrylamide derivatives could also be easily extracted by ethyl acetate from the aqueous phase just like acrylamide itself Meanwhile, the adding volume and frequency of ethyl acetate should be additionally optimized in order to achieve an excellent recovery of the analyte Therefore, recovery rates for five different kinds of extracting approaches were tested, i.e liquid–liquid extraction by (i) ml; (ii) × ml; (iii) × ml; (iv) × ml; (v) × ml of ethyl acetate Corresponding results were shown in Fig 2, and Fig The recovery change of acrylamide derivative (2-BPA) by the use of different volumes of ethyl acetate during the liquid–liquid extraction step The recovery was calculated as the ratio of 2-BPA content in the standard system after and before liquid–liquid extraction n × (ml) means the extraction step was performed using ethyl acetate (4 ml each, n times) Data were expressed as mean ± SD (n = 3) Y Zhang et al / J Chromatogr A 1116 (2006) 209–216 213 Table The retention time (tR ) of 2-bromopropenamide (2-BPA) and numbers of theoretical plates (N) calculated for the separation of acrylamide and performed on different GC columns Column Dimensions Solid phase tR (min) N DB-5 DB-23 DB-WAX HP-5 HP-35 HP-INNOWax HP-FFAP 30 m × Ø0.25 mm, 0.25 ␮m 30 m × Ø0.25 mm, 0.25 ␮m 30 m × Ø0.25 mm, 0.25 ␮m 30 m × Ø0.25 mm, 0.25 ␮m 30 m × Ø0.25 mm, 0.32 ␮m 30 m × Ø0.25 mm, 0.25 ␮m 30 m × Ø0.53 mm, 0.25 ␮m (5%-Phenyl)-methylpolysiloxane (50%-Cyanopropyl)-methylpolysiloxane Polyethylene glycol (PEG) (5%-Phenyl)-methylpolysiloxane (35%-Phenyl)-methylpolysiloxane Polethylene glycol (PEG) Nitroterephthalic acid modified polyethylene glycol 7.1 8.9 11.3 6.6 8.6 12.3 11.6 3103 6020 64164 1862 6556 83815 71652 × ml of ethyl acetate was chosen as the optimal liquid–liquid extracting approach according to the better recovery rate and solvent cost because no considerable recovery improvement presented for nearly all of standards and sample matrixes using × ml and × ml of ethyl acetate compared with the extraction volume of × ml After dryness over sodium sulfate (see Section 2), the analyte extract was clean enough for direct injection into the GC–ECD system 3.4 Optimization of GC column and detection sensitivity of ECD Acrylamide is regarded as strong polar molecule with poor retention and bad peak shape in routine non-polar or weak polar GC columns, such as DB-5, DB-23, HP-5, etc Therefore, GC-based methods with ECD seem not to be capable of analyzing the acrylamide level quantitatively in fried foods on the basis of previous published studies because of the above disadvantages Although derivatization of acrylamide could prolong the retention time of the analyte to a certain extent, the choice of GC columns needs to be optimized and the detection precision and sensitivity of ECD should be examined under such conditions By contrast, Găokmen et al [22] optimized the type of chromatographic columns and demonstrated that the retention of acrylamide could be improved by both hydrophilic and hydrophobic interaction chromatography by avoiding organic modifiers like acetonitrile and methanol in the aqueous mobile phase In the present study as for GC method, some different types of GC capillary columns including DB-5, DB-23, DB-WAX, HP-5, HP-35, HP-INNOWax, and HP-FFAP were tested for their ability to retain bromination products of acrylamide and elute peak responses The retention time and numbers of theoretical plates calculated for the separation of acrylamide and performed on different GC columns have been shown in Table Using nitrogen as the carrier gas at a flow rate of ml min−1 , three columns (DB-23, DB-WAX, and HPINNOWax) among the tested columns were found to have the numbers of theoretical plates exceeding 20,000 The greatly improved elution efficiency using these columns is probably due to their high polarity characteristics of solid phase while other columns have a relatively non- or mid-polarity solid phase Furthermore, considering the retention of bromination products of acrylamide, HP-INNOWax was regarded as the most appropriate capillary column for the quantification and separation of acrylamide By contrast, many co-extractives could be detected during the separation of acrylamide due to no cleanup procedures were performed However, when the GC analysis for the sample extracts was performed on HP-INNOWax capillary column (see Fig 3), acrylamide response could be very well detached from such unidentified co-extractives under the optimized chromatographic conditions, while the others especially DB-5, DB-23, HP-5, and HP-35 column failed to separate Fig GC–ECD chromatogram of acrylamide derivative (2-BPA) in potato chips Retention time of 2-BPA: 12.266 214 Y Zhang et al / J Chromatogr A 1116 (2006) 209–216 acrylamide from these interferences because of excessive peak broadening The electron capture detector is a halogen-sensitive detector widely used for the analysis of pesticide residues in water [31], wildlife whole blood [32], and wildlife food resources [33] GC–ECD method is more sensitive than the HPLC–ESI–MS/MS method when employing standard solutions according to our previous studies about HPLC–ESI–MS/MS analysis [15] Both the limit of detection and the limit of quantification were evaluated by the statistical software of Chemstation for 3D (Agilent, Palo Alto, CA, USA) for GC–ECD The matrix for LOD and LOQ calculations was based on step-by-step dilution of the representative conventional fried sample preparation Compared with results of LC–MS/MS analysis published in peer-review journals [11], the LOD of GC–ECD method is lower (0.1 ␮g kg−1 ) but its LOQ value is higher (3 ␮g kg−1 ) than the former technique when employing standard solutions; the contrary results of LOD and LOQ were found for quantitative analysis of acrylamide in sample extracts owing to high-resolution potential of MS/MS and relative instability of ECD However, it is enough to accurately quantify the acrylamide contaminant in conventional fried foods For instance, acrylamide contaminant was successfully detected in potato chips by ECD as shown in Fig By contrast, the calibration sample extracts were done by the external method with a concentration sequence of 0.5, 1, 5, 10, 50, 100, and 125 ng ml−1 after derivatization with KBrO3 and KBr, and liquid–liquid extraction with ethyl acetate (see the details in Section 2.3) Results showed that the response of ECD was linearly changed with the concentration of acrylamide (y = 417,089x − 473, n = 3) Excellent linearity was obtained with typical values for the correlation coefficient (R2 ) between 0.999 and 1.000 The calibration range could be extended to higher amounts, but this was not routinely done because corresponding linear ranges for GC–ECD analysis were sufficiently wide to measure acrylamide levels from current sample matrixes 3.5 Method validation Instrumental methods using standard solutions and complete methods using spiked and non-spiked sample matrixes were validated for GC–ECD of acrylamide bromination derivatives For Table Repeatability test of acrylamide in conventional fried foods No Sample name Acrylamide level (␮g kg−1 ) U (%)a Potato crisps T3011b Potato crisps Potato chips Potato chips Fried chicken wings Fried chicken wings 1381 2650 453 465 164 225 5.7 6.6 2.6 6.2 5.9 6.0 a b Data were expressed as mean and U (n = 5) U, measurement uncertainty Certified reference test material from Central Science Laboratory qualitative purposes, the method was evaluated by taking into account the precision of retention time, the interference of coelution, and peak purity of the analyte A high repeatability of the retention time was obtained with RSD values lower than 5% for both standards (non-labeled and labeled acrylamide) and different food matrixes For quantitative purposes, the method was validated by defining the repeatability, precision, and recovery The repeatability of the method was estimated for determination of acrylamide from some representative conventional fried foods including potato crisps (one were purchased from the supermarket and the other, i.e FAPAS T3011 potato crisps, were obtained from Central Science Laboratory, UK), potato chips, and fried chicken wings Good repeatability was obtained for six representative above-mentioned samples by GC–ECD analysis, and the measurement uncertainty (U) was estimated for the GC-based technique according to the formula, i.e U (%) = × CVrepeatability (%), proved by previous study (Table 2, n = 5) [34] The precision of the method was determined by calculating the relative standard deviations (RSD) of the replicate measurements Table showed the results of a 5-day precision study in which portions of three levels of acrylamide in three conventional fried products were repeatedly analyzed three times each day for the inter-day precision study while the other portions were repeatedly analyzed six times in Day for the intra-day precision study Accuracy for acrylamide by GC–ECD was excellent with the values for the RSD of each data set ranged between 0.8 and 5.9% (Table 3) Recovery rates were demonstrated in three tests employing the method of standard addition ranging from 150 to 1000 ␮g kg−1 The matrixes included three kinds of special Table Intra- and inter-day precision study of acrylamide levels measured in three commercial conventional fried products by GC–ECD Potato crisps Potato chips Fried chicken wings Intra-day precision (n = 6) Acrylamide level (mean ± SD, ␮g kg−1 ) Day 1360 ± 80 458 ± 131 ± Inter-day precision (n = 3) Acrylamide level (mean ± SD, ␮g kg−1 ) Day Day Day Day 1344 ± 37 1364 ± 12 1356 ± 11 1347 ± 15 457 ± 12 448 ± 12 462 ± 444 ± 130 ± 131 ± 127 ± 129 ± Total mean (␮g kg−1 ) SD (␮g kg−1 ) 1355 46 454 10 130 Y Zhang et al / J Chromatogr A 1116 (2006) 209–216 215 Table Spiking recovery test of samples by GC–ECDa Sample (no.) (␮g kg−1 ) Before addition Added amount (␮g kg−1 ) After addition (␮g kg−1 ) Recovery (%) Fried chicken wings (I) Potato chips (II) Potato crisps (III) 143 ± 150 275 ± 87 ± 466 ± 19 500 931 ± 13 93 ± 1415 ± 51 1000 2384 ± 46 97 ± a The concentration data of acrylamide were shown as means ± SD (n = 4).Three representative samples spiked at 150, 500, and 1000 ␮g kg−1 of acrylamide standard solution were regarded as low, intermediate, and high level spiked food matrix, respectively I, II, samples purchased from two major Western food suppliers III, certified reference test material (FAPAS T3011 potato crisps) samples which represented the low, intermediate, and high levels of acrylamide containing foods from local supermarkets and FAPAS certified test material Acrylamide standard was added to each sample at a certain level, which was close to the determined acrylamide level of corresponding sample Total acrylamide derivative (2-BPA) was then determined for each level Subtracting the incurred residue from the total amount of acrylamide derivative found indicated the spiking recovery for each sample Good recoveries were achieved for the validated GC–ECD method, confirming the reliability of the methods including the bromination derivatization prior to GC analysis The mean percentage recoveries exceeded 85% for all spiking levels for conventional fried foods (Table 4) Six independent measurements of the certified reference material averaged to an acrylamide concentration of 1381 ␮g kg−1 (satisfactory assigned value: 1404 ␮g kg−1 ) with a RSD value of 4.2% Considering all of the above data for method validation test, the GC method established in the present study and sample pretreatment procedures employed in the present work can be regarded as selective, precise, and robust Conclusion The present study developed a GC–ECD method for identification and quantification of acrylamide in conventional fried complex food matrixes such as potato crisps, potato chips, and fried chicken wings To our best knowledge, this is the first GC–ECD method dealing with contaminant analysis of acrylamide and the first modified method employing potassium bromate and potassium bromide as derivatization reagents for acrylamide derivatization, i.e 2-BPA, prior to GC separation The method presented in this study is sensitive enough for the analysis of acrylamide in conventional fried foods with the LOD and LOQ value at 0.1 and ␮g kg−1 , respectively Meanwhile, the present GC–ECD analytical method requires a relatively low-cost instrumentation compared with LC and GC with tandem MS methods already published in many journals, and also can be applied for the investigation of acrylamide contaminant in local characteristic heat-treated foods by many laboratories world wide easily Furthermore, the GC–ECD method showed that no clean-up steps of acrylamide derivative is necessary prior to injection and was slightly more sensitive than the tandem MS/MS-based methods Extraction and derivatization proved to be the crucial steps in sample preparation, and further optimization of these processes should be performed For instance, a convenient and fast pretreatment procedure should be optimized in order to satisfy the investigation of hundreds of samples Such procedure mainly depends on the optimization of extraction solvents and corresponding parameters Delatour et al [24] achieved the visual protein precipitation step within