Acrylamide in Foods: Chemistry and Analysis. A Review
- First Online:
- Cite this article as:
- Keramat, J., LeBail, A., Prost, C. et al. Food Bioprocess Technol (2011) 4: 340. doi:10.1007/s11947-010-0470-x
- 936 Views
Acrylamide is a potential cause of a wide spectrum of toxic effects and is classified as probably “carcinogenic in humans”. The discovery of acrylamide in human foods has given rise to extensive studies exploring its formation mechanisms and levels of exposure and has spurred search into suitable analytical procedures for its determination in foodstuffs. However, the exact chemical mechanisms governing acrylamide formation are not yet known and cheap, convenient, and rapid screening methods are still to be developed. Acrylamide in food is produced by heat-induced reactions between the amino group of asparagine and the carbonyl group of reducing sugars along with thermal treatment of early Maillard reaction products (N-glycosides). Similarly, the decarboxylated Schiff base and decarboxylated Amadori compounds of asparagine as well as the Strecker aldehyde have been proposed as direct precursors and intermediates of acrylamide. Corresponding chromatographic methods are used to determine various structural groups present in Maillard reaction model systems. Gas chromatography-mass spectrometry and liquid chromatography with tandem mass spectrometry analysis are both acknowledged as the main, useful, and authoritative methods for acrylamide determination. This review is an attempt to summarize the state-of-the-art knowledge of acrylamide chemistry, formation mechanisms, and analytical methods. Special attention is given to comparison of different chromatographic techniques, particularly the novel, simple, and low-cost methods of its determination.
KeywordsAcrylamide chemistryMechanism of acrylamide formationAcrylamide analysis
Acrylamide content of various food products
Number of samplesa
Mean concentration, μg/kg
Reported maximum, μg/kg
Cereal and cereal-based products
Cereal and pasta, raw and boiled
Cereal and pasta processed (toasted, fried, grilled)
Cereal-based processed products, all
Bread and rolls
Pastry and biscuits
Fish and seafood (including, e.g., breaded, fried, baked)*
Meat and offal (including, e.g., coated, cooked, fried)
Milk and milk products
Nuts and oilseeds
Root and tubers
Potato crisps (US=chips)
Potato chips (US=French fries)
Potato chips, croquettes (frozen, not ready to serve)
Stimulated and analog
Coffee (brewed), ready-to-drink
Coffee (ground, instant, or roasted, not brewed)
Green tea (roasted)
Sugar and honey (mainly chocolate)
Raw, boiled and canned
Processed (toasted, baked, fried, grilled)
Fruits, dried, fried, processed
Alcoholic beverages (beer, gin, wine)
Condiments and sauces
Baby food (canned, jarred)
Baby food (dry powder)
Baby food (biscuits, rusks, etc.)
Contribution of food groups to acrylamide exposure (%) in different countries for different age groups (Heatox Report 2007)
The Netherlands (1–97 years)
The Netherlands (1–6 years)
Sweden (18–74 years)
Germany (15–18 years)
Norway (16–79 years)
Belgium (13–18 years)
Potato products, fried
Chocolate and chocolate-spread
Acrylamide is a potential cause of a wide spectrum of toxic effects (Eriksson 2005; IARC 1994; European Union Risk Assessment Report 2002; Manson et al. 2005), including neurotoxic effects as has been observed in humans. Also, acrylamide has been found to be carcinogenic in animals, increasing incidences of a number of benign and malignant tumors identified in a variety of organs (for example thyroid, adrenals; FAO/WHO 2004). Several observations led to the hypothesis that heating of food could be an important source of acrylamide exposure to humans, if the heating/frying is done with a frying pan, in an oven, or by microwave heating; but no acrylamide has been detected in boiled food products (Törnqvist 2005).
There are two established legal limits for acrylamide. One concerns drinking water (WHO 1996; EEC 1998), and the other involves the migration of acrylamide from packaging materials into food (EEC 1992). The latter is defined not to be detectable within a limit of detection (LOD) of 10 μg acrylamide in 1 kg of food, while a daily intake of some tens of micrograms can be expected depending on dietary habits. This alarmed food producers as well as food control authorities. Valuable information about acrylamide and its toxicological properties have been recently summarized in the report released by the Scientific Committee on Food (SCF) (2002) and the Heatox Report (2007).
Most studies (Zhang et al. 2005; Vattem and Shetty 2003; Schabacker et al. 2004) have focused on the mechanisms of acrylamide formation in heat-treated foods. Since these mechanisms have not been completely understood, it is not yet possible to determine an effective pathway for reducing its occurrence in different heat processing technologies by controlling critical steps of food processing.
Acrylamide content in foods is defined as the net amount of acrylamide, i.e., what remains after formation and degradation. It has been shown that prolonged storage and increase of temperature and heating time enhance the acrylamide formation until it attains a maximum point of acrylamide formation, then acrylamide degrades, and the net amount of it decreases. The type of reaction responsible for the degradation is still unclear (Biedermann et al. 2002b; Biedermann et al. 2002c; Weibhaar 2004; Hoenicke and Gatermann 2004; Hoenicke and Gatermann 2005; Delatour et al. 2004; Eriksson and Karlsson 2005). However, the stability of acrylamide at 190 °C has been evaluated by pyrolysis of 13C3-labeled acrylamide in a glucose/asparagine model reaction. At this temperature, acrylamide mainly occurs in the polymeric form with CH2CHCONH2 as the monomeric unit. Also, NMR experiments have indicated that polymerization of acrylamide can easily occur. On the other hand, acrylamide may also react with soft nucleophiles according to the hard and soft acid base theory and can, therefore, be consumed in Michael type addition reactions (Stadler et al. 2004).
Rather than acrylamide, N-methylacrylamide and 3-buteneamide are the new compounds formed during baking processes. These compounds are suspected toxicants in food products.
It is the main objective of most research efforts in the field to determine how it is possibly formed or decomposed and how it can be accurately measured. Therefore, this review presents, in two separate sections, the recent hypotheses put forward and the factors involved in acrylamide formation mechanisms and the methods of analysis used for its determination.
Mechanisms of Acrylamide Formation
Becalski et al. (2003) developed two models. The first one involved a mixture of the following six amino acids: asparagine, aspartic acid, glutamine, glutamic acid, valine, and lysine along with glucose as a reducing sugar (Yang et al. 1999). The second model was a simplified version thereof which consisted only of asparagine and glucose. These researchers found that acrylamide was not principally formed from precursors (especially acrolein) present in the oil itself (Becalski et al. 2003). They also reported that, in model reactions consisting of a mixture of six amino acids, the yield of acrylamide at 175 °C was 3300 ng of acrylamide per about 23 mg of each amino acid, but, it was not formed in a mixture of aspartic acid, glutamine, and lysine under identical conditions. Nor was it detected in samples of amino acids and glucose heated at 120 °C or 140 °C. They concluded that the presence of asparagines and glucose at the right temperature should have important roles in the formation of acrylamide and that its formation could be expected at heating temperatures above 175 °C for more than 10 min.
In a similar study, Mottram et al. (2002) reported that a significant quantity of acrylamide (221 mg per mol of amino acid) formed when an equimolar mixture of asparagine and glucose reacted at 185 °C in the phosphate buffer. Their findings also confirmed that the presence of asparagine and glucose was critical for the formation of acrylamide.
The presence of asparagine with glucose or 2,3-butanedione (one of several dicarbonyl compounds formed in the Maillard reaction) causes significant amounts of acrylamide to form in dry products, but only trace amounts would form when asparagine is replaced with other amino acids. Heating asparagine on its own at 185 °C does not produce acrylamide, confirming the requirement for the dicarbonyl reactant to be present and the Strecker degradation to occur (Mottram et al. 2002). However, replacing glucose with other carbohydrates (d-fructose, d-galactose, lactose, and sucrose) concurrent with replacement of asparagine led to a significant amount of acrylamide to be released (Stadler et al. 2002).
The above studies revealed that the type and quantity of precursors play important roles in acrylamide formation. Most amino acids other than asparagines do not take part in the reactions leading to acrylamide formation. Also, the presence of both the amino group (from asparagine) and the carbonyl group (from reducing sugar) or the dicarbonyl groups (from Maillard reactions) is highly crucial for acrylamide formation. Temperature and solvent also have important effects on acrylamide formation that must be duly considered.
The above statement is applicable to such heat-treated, plant-based foods as cereals and potato which are rich in asparagine as a free amino acid (Magnin 1964; asparagines accounting for 40% and 14% of their total free amino acids, respectively; Borodin 1925). These foodstuffs are the most likely ones to produce the highest amount of acrylamide in the foods prepared from them.
The early Maillard intermediates such as N-glycosylasparagines yield more acrylamide under milder reaction conditions than the binary mixtures of the precursors. The decarboxylated Schiff base and decarboxylated Amadori compounds of asparagine have been proposed to be direct Maillard precursors of acrylamide (Yaylayan et al. 2003; Zyzak et al. 2003). Also, the Strecker aldehyde has been detected as another direct intermediate of acrylamide (Mottram et al. 2002). The pathway of acrylamide formation in a glucose/asparagine system occurs prior to the Amadori rearrangement and, consequently, the amount of acrylamide released from N-glycosyl asparagine is about 20 times higher than from the Amadori compound. The protection of the hydroxyl groups in the sugar ring and in the carboxyl groups does not affect acrylamide formation from Amadori compound of asparagine. This is because the Amadori compound is not easily decarboxylated. These compounds are the first stable intermediates generated as a result of the early Maillard reaction leading to 1- and 3-deoxyosones, which further decompose to generate color and flavor compounds (Ledl and Schleicher 1990). However, in low-moisture systems limiting the reversibility of the initial step, the first stable intermediates are the N-glycosyl compounds, which mainly rearrange via the corresponding Schiff base to the Amadori compound, the 1-deoxyfructosyl derivative of the amino acid, which is not a favored Maillard intermediate to generate acrylamide. The Schiff base may alternatively decarboxylate to the intermediary azomethine yield, which after tautomerization leads to the decarboxylated Amadori compound. The vinylogous compounds are then released along with the corresponding aminoketone by a β-elimination reaction and cleavage of the carbon–nitrogen covalent bond (Yaylayan et al. 2003; Zyzak et al. 2003).
Decarboxylation of the Schiff base may proceed via the zwitter ionic form which is more probable than the classical Strecker degradation mechanism (Grigg et al. 1988; Grigg and Thianpatanagul 1984; Schönberg and Moubacher 1952). Zyzak et al. (2003) reported evidence for the decarboxylated Schiff base of asparagine, which may also represent the decarboxylated Amadori compound. They suggested acrylamide to be formed directly from this compound. Yaylayan et al. (2003) suggested a decarboxylated Maillard intermediate as a direct precursor to acrylamide. The proposed pathway is based on Intramolecular cyclization of the Schiff base to the oxazolidine-5-one derivative. Such oxazolidine-5-ones have been reported to easily decarboxylate, thus giving rise to stable azomethine yields, which after tautomerization reacts as a direct precursor to acrylamide (Manini et al. 2001).
Investigations published in 2002 have revealed that the Maillard reaction is one major reaction pathway, particularly in the presence of asparagine, which directly provides the backbone of the acrylamide molecule (Biedermann et al. 2002b; Becalski et al. 2003; Mottram et al. 2002; Stadler et al. 2002; Sanders et al. 2002; Weibhaar and Gutsche 2002). However, other reaction pathways have also been suggested, such as acrolein released by oxidative lipid degradation leading to acrylic acid, which can react with ammonia to give acrylamide. Acrylic acid can also be generated from aspartic acid by the Maillard reaction (Zyzak et al. 2003; Gertz and Klostermann 2002; Stadler et al. 2003).
There are basically two major hypotheses put forward so far regarding the formation of acrylamide from asparagine by the Maillard reaction. Mottram et al. (2002) suggested that α-dicarbonyls were necessary as coreactants in the Strecker degradation reaction affording the Strecker aldehyde as the precursor to acrylamide. Stadler et al. (2004) proposed glycoconjugates, such as N-glycosides and related compounds formed in the early stage of Maillard reaction, as the key intermediates leading to acrylamide. Yaylayan et al. (2003) and Zyzak et al. (2003) have supported this hypothesis by showing the importance of the Schiff base of asparagine, which corresponds to the dehydrated N-glycosyl compound. The key mechanistic step is decarboxylation of the Schiff base which leads to the formation of Maillard intermediates that can directly release acrylamide (Fig. 3).
The conclusion to be drawn from the above observations is that the main pathways for acrylamide formation can be classified as follows.
Pathways of Acrylamide Formation
Asparagine Route of Acrylamide Formation via Maillard Reaction
One of the major pathways of acrylamide formation is the asparagine route (Gökmen and Palazoğlu 2008). Also, different other routes have been suggested in conjunction with the Maillard reactions system in food products (Mottram et al. 2002; Stadler et al. 2002). Asparagine can principally be converted to acrylamide through thermal decarboxylation and deamination reactions, but the presence of reducing sugars is essential for these reactions to occur. Many carbonyl-containing moieties can enhance a similar transformation (Stadler et al. 2004; Becalski et al. 2003; Yaylayan et al. 2003; Zyzak et al. 2003; Stadler et al. 2003). In model studies, it has been shown that α-hydroxy carbonyls are much more efficient than α-dicarbonyls in converting asparagine to acrylamide (Stadler et al. 2003).
Furthermore, according to the results obtained from model systems, fructose increases the acrylamide content by about two times in comparison with other reducing sugars because its contains two α-hydroxylic groups rather one as is the case with other sugars (Eriksson 2005).
In addition, decarboxylated asparagine (3-aminopropionamide) can generate acrylamide in the absence of reducing sugars. Structural considerations dictated that asparagine alone might be converted thermally into acrylamide through decarboxylation and deamination reactions. However, the main product of the thermal decomposition of asparagine is maleimide, mainly due to the fast intramolecular cyclization reaction that prevents the formation of acrylamide. On the other hand, asparagine (in the presence of reducing sugars) is able to generate acrylamide in addition to maleimide (Yaylayan et al. 2003; Zyzak et al. 2003).
Further evidence to support this pathway to acrylamide production was provided by Becalski et al. (2003). They reported that heating of asparagine-15N-amide hydrate and glucose at 175 °C produces a compound with similar characteristics to those of [15N] acrylamide and the yield of reactants are similar to those obtained with unlabeled asparagine. This is in accordance with the results reported by Stadler et al. (2002) who showed that 98.6% of nitrogen-15 label was incorporated into acrylamide after pyrolysis of 15N-amide-labeled asparagine with glucose while incorporation into acrylamide was observed when 15N-α-amino-labeled asparagine was used in the same reaction. Therefore, it appears that only the amido nitrogen of asparagine is being incorporated into acrylamide in this reaction and that acrylamide is formed through the deamination and decarboxylation of asparagine and the formation of a C–C double bond. The exact nature of the reaction pathway is supposed to be the reaction of asparagine with a carbonyl moiety followed by rearrangement(s) of an intermediate (Becalski et al. 2003).
Alternative Pathways of Acrylamide Formation
Acrolein (1-propenal) is a simple α- or β-unsaturated aldehyde which is supposed to be a probable cytotoxic compound (Casella and Contursi 2004). In food, particularly in oil and fat, an alternative pathway for the formation of acrylamide through acrolein has been proposed as the mechanism via acrylic acid (Gertz and Klostermann 2002; Stadler et al. 2003; Becalski et al. 2002). In fact, recent reports on model reaction systems demonstrate that acrolein, together with asparagine, may generate appreciable levels of acrylamide under certain conditions, suggesting a critical role for acrolein in the formation of acrylamide in lipid-rich foods (Zhang et al. 2005; Yasuhara et al. 2003).
Acrylamide could be produced from oils and nitrogen-containing compounds present in foods. The most plausible scheme includes the transformation of acrolein to acrylic acid and the final reaction of acrylic acid with ammonia, which could potentially be generated by pyrolysis of nitrogen-containing compounds leading to the formation of acrylamide (Becalski et al. 2003).
Aspartic acid, carnosine, and β-alanine can give rise to acrylamide formation through the formation of acrylic acid during their thermal decomposition in combination with available ammonia to convert acrylic acid to acrylamide (Yaylayan et al. 2004; Stadler et al. 2003; Yaylayan et al. 2005; Sohn and Ho 1995).
Aminopropionamide has been identified as an intermediate during acrylamide formation from asparagine. This compound is also formed in reactions between asparagine and pyruvic acid and is a very effective precursor to acrylamide formation (Stadler et al. 2004; Zyzak et al. 2003).
Pyruvic acid can be generated by dehydration and desulfidation of serin and cysteine, respectively. It can then be proposed as the reduction of pyruvic acid into lactic acid, with further dehydration into acrylic acid. Finally, acrylic acid is transformed to acrylamide (Yaylayan et al. 2005; Wnorowski and Yaylayan 2003).
Formation of acrylamide may occur by α-dicarbonyl-assisted Strecker dehydration. Compared with sugar/asparagine mixtures, co-pyrolysis of asparagine with various dicarbonyls such as α-diketones, α-ketoaldehydes, and glyoxal give relatively lower acrylamide concentrations, while acetol generates the highest amount of acrylamide. However, the Strecker alcohol of asparagine (3-hydroxypropanamide) has been found to generate acrylamide by a one-step dehydration reaction, but only at concentrations lower than those reported for sugar/asparagine mixtures. These data suggest that the Strecker aldehyde of asparagine via the alcohol has limited importance in the formation of acrylamide (Stadler et al. 2004).
Benzaldehyde and styrene are formed as volatile compounds during pyrolysis of Amadori compounds. As the latter compound represents the decarboxylated Amadori compound of phenylalanine, acrylamide may be formed from the decarboxylated Amadori compound of asparagine (Stadler et al. 2004).
The sugar-asparagine adducts, i.e., N-glycosylasparagine, generate high amounts of acrylamide, suggesting the early Maillard reaction as a major source of acrylamide (Stadler et al. 2002). Good evidence has been provided that supports the early Maillard reaction as a main reaction pathway involving early decarboxylation of the Schiff base, rearrangement to the resulting Amadori product, and subsequent β-elimination to release acrylamide (Yaylayan et al. 2004).
The early Maillard products such as N-glycosides have been found to be acrylamide precursors in thermal decomposition reactions. On the basis of structural considerations, asparagine or the N-glycosides could be direct precursors to acrylamide under pyrolytic conditions. The amino acid is the carbon source of acrylamide and the formation of the corresponding N-glycoside probably facilitates the decarboxylation step and heterolytic cleavage of the nitrogen–carbon bond to liberate acrylamide (CH2=CHCONH2) upon pyrolysis (Stadler et al. 2002).
In food processing systems that incorporate conditions of high temperature and water loss, N-glycoside formation could be favored. Maximum acrylamide formation by thermal decomposition of early Maillard intermediates has been observed after an incubation period of 1 h at 180 °C (Stadler et al. 2004).
Corresponding Methods for Analysis of Model System Products
Pyrolysis–Gas Chromatography/Mass Spectrometry
Pyrolysis techniques are based on the principle of increasing the amount of heat energy in the system, thus leading to thermal cracking of bands of non-volatile organic compounds. The pyrolysis–gas chromatography/mass spectrometry (Py–GC/MS) technique has a large potential for evaluation and analysis of complex samples, and usually no pre-treatment is needed. Polar compounds such as acrylamide can be derivatized directly on the filament in the pyrolysis chamber (Zhang et al. 2005).
To testify the amide amino acid asparagine, 13C isotope tracer technique is necessarily applied for the mechanistic study of acrylamide. Py–GC/MS is also a convenient method to perform experiments with labeled reactants. Position and label distribution in the common products in the same (aqueous and pyrolytic) model system are identical. Thus, the mechanisms of acrylamide formation under both conditions are similar (Yaylayan 1999; Wnorowski and Yaylayan 2000). Yaylayan et al. (2003) identified niacinamide by Py–GC/MS assays with 13C-labeled glucose in the reaction mixture of the asparagine/glucose model system. Niacinamide can form only from the decarboxylated Amadori product of asparagine with glyceraldehydes through cyclization and dehydration reactions. During Py–GC/MS assays, it is easier for researchers to confirm the chemical structure of each intermediate or final product by using the mass spectrometric data than the US National Institute of Science and Technology (NIST) library in the Py–GC/MS software (Keyhani and Yaylayan 1996).
Fourier Transform Infrared Analysis
Fourier transform infrared (FT-IR) is a powerful tool for identifying types of chemical bands in a molecule by producing an infrared absorption spectrum that is like a molecular “finger-print”. FT-IR is also the most useful method for identifying chemicals present in organic samples.
When glucose/asparagine model reactions are performed, the facile formation of a decarboxylated asparagine Amadori product is detected by IR spectral analysis of a mixture of asparagine and glyceraldehyde in methanol. Also, the carbonyl band of the Amadori product is indicated by the appearance of the absorption band at 1,737 cm-1. As the reactions continue, bands at 1,493 and 1,580 cm-1 appear that indicate the decarboxylation reaction and formation of a decarboxylated Amadori product. When the temperature of the mixture is raised to 180 °C, the data show the disappearance of the amide I (1,674 cm-1) and amide II (1,615 cm-1) absorption bands, indicating cleavage of the asparagine moiety (Yaylayan et al. 1999a; Yaylayan et al. 1999b).
Other Methods of Analysis
Besides the chromatographic (Py–GC/MS) and spectroscopic (FT-IR) techniques, other methods including high-performance liquid chromatography with ultraviolet diode array detection, liquid chromatography with tandem mass spectrometry (LC-MS/MS) (Robert et al. 2004; Terada and Tamura 2003; Granvogl et al. 2004; Taubert et al. 2004), and time of flight mass spectrometry (Sanders et al. 2002) also play an important role in the elucidation of the generating mechanism of acrylamide during heat processing. The critical tools of these three techniques are UV scan spectrograms, multiple reaction-monitoring mode (MRM), quantitative analysis and precise mass assignment, respectively (Zhang et al. 2005; Hamdan et al. 2001).
Methods of Acrylamide Analysis
Recent Analytical Methods
The discovery of acrylamide in human foods led to surveys exploring the levels of this potentially hazardous chemical and spurred search into suitable analytical procedures for its determination in foodstuffs (IARC 1994; Tareke et al. 2002; Becalski et al. 2003; Zhu et al. 2008; Ahn et al. 2002; Rosén and Hellenäs 2002; Tareke et al. 2000). The potential presence of acrylamide in foods was initially investigated by employing derivatization of acrylamide (bromination of a double bond) and subsequent gas chromatography-mass spectrometry (GC-MS) detection (Andrawes et al. 1987; Castle 1993; Castle et al. 1991; United States Environmental Protection Agency 1996). Later, several groups reported using the same techniques, but without derivatization (Biedermann et al. 2002b; Rothweiler and Prest 2003; Tateo and Bononi 2003). An LC-MS/MS-based method was then developed, and soon, a few more variants of that procedure appeared in the literature (Rosén and Hellenäs 2002; Ono et al. 2003; Swiss Federal Office of Public Health 2002; Takatsuki et al. 2003; US Food and Drug Administration 2003). Although MS is chosen as the main technique for GC- and High Performance Liquid Chromatography HPLC-based analysis, there is still a need to develop a reliable, sensitive, rapid, and low-cost analytical method for the determination of acrylamide without using MS. Examples include the US EPA Method 8032A that uses liquid extraction and Gas Chromatography-Electron Captured Detector (GC-ECD) for determinations in water and a method by the German Health Agency (BGVV) (Results of a BGVV Information Seminar 2002) that uses HPLC with UV detection for migration analysis of acrylamide from food packing materials (Dionex 2004). However, those methods are not easily applied for determination of acrylamide in a broad range of foodstuffs.
Acrylamide does not show any specific wavelength absorption maxima (λmax). Therefore, by performing the same type of bromination as for GC analysis, it is possible to achieve a detection limit down to 4 μg/L by using LC, low UV-detection (Brown and Rhead 1979; Brown et al. 1982).
Today, the detection limit is decreased to 0.5 μg/L by direct injection on a LC-MS of the brominated derivate (Cavalli et al. 2004). Over the recent years, many more papers and reviews have been published about the occurrence and analytical methods of acrylamide in heated foods (Wenzl et al. 2003, Zhang et al. 2005; Castle and Eriksson 2005; Eberhart et al. 2005; Kim et al. 2007; Pittet et al. 2004; Ren et al. 2006).
Becalski et al. (2003) determined the levels and sources of acrylamide in the Canadian food supply and developed a LC-MS/MS method. This method incorporated several purification steps and might be useful for determination of acrylamide by detection of its bromo derivative as a relatively clean extract is obtained (Castle 1993). Gökmen et al. (2005) developed a sensitive reversed-phase HPLC-DAD method for acrylamide analysis in potato-based processed foods. Although a rapid and convenient measurement was successfully achieved, the method required a relatively complex sample pre-treatment including extraction of acrylamide with methanol, purification with Carrez I, and Carrez II solutions, evaporation and solvent change to water, and clean-up with an Oasis HLB solid phase extraction (SPE) cartridge. It was also found that the sensitivity of the HPLC method was lower than those of GC- and MS-based methods (Zhu et al. 2008).
Analysis for acrylamide by bromination and GC determination is relatively advanced and is nowadays used for acrylamide determination in heated foods (Castle 1993; Castle et al. 1991; United States Environmental Protection Agency 1996; Bologna et al. 1999; Poole 1981). Recently, Zhang et al. (2006, 2007) developed a GC-ECD method for identification and quantification of acrylamide in fried foods such as potato crisps, potato chips, and fried chicken wings. The method showed a lower limit of detection compared with MS-based methods. Also, Zhu et al. (2008) developed a low-cost, convenient, sensitive, and accurate quantitative method for the determination of low-level acrylamide in heat-processed starchy foods by GC-ECD using the standard addition method. Hamlet et al. (2004) developed a rapid, sensitive, and selective analysis of acrylamide in cereal products using bromination and GC-MS/MS. They also studied the kinetics of acrylamide bromination to develop and apply to MS/MS for selective detection with minimal sample clean-up. These authors found that using standard solutions of 2,3-dibromopropionamide (2,3-DBPA) deteriorated the chromatographic performance due to the build-up of co-extracted materials. Therefore, they recommended deliberate conversion of 2,3-DBPA to 2-bromopropenamide (2-BPA) prior to injection on the GC column. In most these methods, an aqueous extraction and clean-up procedure is based on using a derivatization step to form the brominated acrylamide adduct followed by GC/MS determination (Rosén and Hellenäs 2002; Tareke et al. 2000; Castle 1993; United States Environmental Protection Agency 1996; US Food and Drug Administration 2003).
Brandl et al. (2002) developed a rapid and convenient method for the sensitive determination of acrylamide suitable for different kinds of starch-rich foodstuff, such as potato chips, cereals, biscuits, cookies, crisp bread, etc. In contrast to the aqueous extraction of acrylamide from foods, they applied an organic solvent extraction method to avoid co-extraction of interfering food matrix components, especially starch; thus, they significantly simplified the clean-up procedure. An aqueous back-extraction of the primary organic phase prior to the clean-up procedure enabled them to achieve a selective extraction of the target analyte. The method of analysis used for the resulting aqueous phase was liquid chromatography coupled with tandem mass spectrometry.
A method consisting of a fast, automated extraction using accelerated solvent extraction (ASE) has been presented by Dionex (2004). According to this method, samples are extracted in 20 min using pure water with 10 mM formic acid or acetonitrile. The extracts are directly analyzed by ion chromatography (IC) using a 4-mm ion-exclusion column and both UV and MS detection. Compared with conventional reversed phase columns, using this column allows for the separation of acrylamide from the many co-extractable compounds present in food samples. The advantages of the method include simplicity, speed of analysis, and a degree of automation that allows the analysis of a large number of samples with minimal labor (Richter et al. 1996).
Chromatographic-based methods for the determination of acrylamide in food products
Extraction and clean-up
Detector LOD and LOQ
Liquid chromatography methods
Brandl et al. 2002
Various food products
Accelerated sample extraction using dichloromethane containing 2% ethanol, 2 g sample+IS+200 μl water in five cycles (10 min each at 80 °C, 100 bar), mixing of combined extract with 5 ml water, using aliquot of aqueous extract
Synergy Polar-RP Column, 150 × 3 mm
Mobile phase: water containing 0.1% acetic acid; 0.5 ml/min; column temperature, 40 °C
LC-MS/MS; LOD, 10 ng/ml
Sheath gas, 60 units; auxiliary gas, 10 units; corona current, 5 μA; vaporize temperature, 350 °C; capillary temperature, 150 °C; m/z transitions (collision energy); AA, 75 > 55; IS, 75 > 58
Rosén and Hellenäs 2002
Mashed potato, rye flour, crisp bread, potato crisps
2–4 g sample+IS+40 ml water, homogenization (2 min at 9,500 min-1), centrifugation (3600×g, 10 min), extra-centrifugation for potato chips (10 min at 16,800) after precipitation by freezing, pretreatment of SPE with 1 ml acetonitrile and washing with water, filtration through 0.22 μm
Hypercarb HPLC column, 50 mm × 2.1 mm
Mobile phase, methanol/water = 20/80 (gradient); 0.4 mL/min; Inj, 10μL
LC–MS/MS; LOD, <10 μg/kg; LOQ<30 μg/kg
Capillary voltage, 2 kV; cone voltage, 20 V; source temperature, 125 °C; desolvation temperature, 400 °C; m/z transitions (collision energy); AA, 72 > 72 (0 eV), 55 (9 eV), 54 (16 eV), 44 (20 eV), 27 (14 eV); IS, 75 > 58 (9 eV)
Ono et al. 2003
Various food products
50 g sample+IS+300–400 ml water, homogenization, centrifugation (20 min at 48,000×g), freezing and melting of supernatant, centrifugation (10 min at 21,700×g), fractionation of 0.5–2 ml supernatant on SPE cartridge, fraction collected and centrifugation (10 min at 27,000×g), filtration of supernatant through 0.22 μm syringe filters, centrifugation of filtration with cut-off of 3 kDa (50 min at 14000×g)
Atlantis dC18 column, 150 mm × 2.1 mm
Mobile phase: methanol/water = 10/90; 0.1 mL/min; run time, 10.2 min; Inj, 2 μL
LC–MS/MS LOD, 0.2 ng/mL LOQ, 0.8 ng/mL
Ion spray voltage, 5.2 kV; turbo gas temperature, 450 °C; m/z transitions (collision energy); AA, 72 > 55 (18 eV); IS, 75 > 58 (18 eV)
Roach et al. 2003
Cereal, bread crumb, potato chips, coffee
1 g sample+IS+ 9 ml water, mixing for 20 min, centrifugation (15 min at 9,000 rpm), centrifugation of 5-ml aliquot in spin filtration tube (2–4 min at 9,000 rpm), pretreatment of Oasis HLB SPE cartridges (3.5 mL of methanol followed by 3.5 mL of Water), extract loaded (1.5 ml) and collected
Synergi Hydro-RP 80 A column, 250 mm × 2 mm
Mobile phase, 0.5% methanol/0.1% acetic acid in water; 0.2 mL/min; run-time, 10 min; Inj, 20μL
LC–ESI-MS/MS; LOD, 10 μg/kg
Capillary voltage, 4.1 kV; cone voltage, 20 V; source temperature, 120 °C; desolvation temperature, 250 °C; m/z transitions (collision energy); AA, 72 > 72 (5 eV), 55 (10 eV), 27 (19 eV); IS, 75 > 75 (5 eV), 58 (10 eV), 29 (19 eV)
Becalski et al. 2003
Potato chips, potato crisps, cereals, bread, coffee
16 g sample+IS+80 ml water+ 10 ml dichloromethane, mixing (15 min), centrifugation (2 h at 24,000×g), centrifugation of 10 ml supernatant (4 h at 4,000×g) in 5 kDa centrifuge filter, passing 5 ml of filtrate through Oasis MAX cartridge connected with tandem with Oasis MCX cartridge, loading the elution onto the preconditioned ENVl-carb cartridge, discarding first 1 ml and collecting remaining (f1), washing cartridge with 1 ml water (f2) and 1.5 ml 10% methanol (f3), analyzing f1, f2, f3
[13C3]-acrylamide or [D3]acrylamide
Hypercarb column, 50 × 2.1 mm
Mobile phase, 15% methanol in 1 mM ammonium formate; 0.175 ml/min; Inj, 5–10 μl; column temperature, 28 °C
MS/MS; LOD, ~6 μg/kg
Ionization mode, positive; desolvation temperature, 250 °C; source temperature, 120 °C; desolvation gas flow, 525 L/h; cone gas flow, 50 L/h; collision gas pressure, 2.6 × 10-3 mbar; ion energy, 10 V ; MRM: dwell time; 0.3 s; cone voltage, 34 V; mass span, 0.1 Da; interchannel energy, 0.05–0.1 s; m/z transitions (collision energy); AA, 72 > 55 (11 eV), 72 > 54 (11 eV), 72 > 44 (14 eV), 72 > 27 (16 eV); IS, 75 > 58 (11 ev)
Riediker and Stadler 2003
Breakfast cereal, crackers
Extraction with water, homogenization using a dispersing tool, centrifugation, mixed with acetonitrile to precipitate co-extractives, acetonitrile evaporation, preconditioning with methanol and water (isolute multimode, 2, 2 × 2 mL; Accubond II SCX 1, 1 mL), residual water removal, extract (2 mL) loaded and collected, collected (1 mL) extract charged onto the latter cartridge, effluent collected, filtration through a syringe filter unit
Shodex RSpak DE-613 polymethacrylate gel column, 150 × 6 mm
Mobile phase, 0.01% aqueous formic acid/methanol = 6/4, 0.75 mL/min split to 0.35 mL after the LC column using a PEEKb; T-piece, run time, 12 min; Inj, 50 μL
LC–ESI-MS/MS; LOQ, 45
Capillary voltage, 3.1 kV; cone voltage, 22 V; source temperature, 100 °C; desolvation temperature, 350 °C; m/z transitions (collision energy); AA, 72 > 55 (11 eV), 54 (20 eV), 27 (20 eV)
Mestdagh et al. 2004
1 g sample + IS+10 ml hexane, 10 min shaking, centrifugation (10 min at 4000 rpm), hexane removal, adding 10 ml Milli-Q water, 20 min shaking, centrifugation (20 min at 4,000 rpm), ultrafiltration through 0.45-μm membrane filter, preconditioning of cartridge (5 ml methanol and 5 ml water), loading on Oasis HLB and Varian Bond Elut Accurate cartridge
Atlantis dC18 column, 150 mm × 2.1 mm
Mobile phase, 92% water (containing 0.1% acetic acid) and 8% water/methanol (35/65, with 0.3% fomaric acid); 0.15 ml/min
m/z transitions (collision energy); AA, 72 > 72 (5 eV), 72 > 55 (10 eV); IS, 75 > 58 (10 eV), 75 > 30 (20 eV)
Hoenicke et al. 2004
Various food products
Weighed into a filter, placed on a Witt'scher pot, equipped with a vacuum pump, defatted by adding iso-hexane (80 mL), spiked with IS, incubation (30 min), extraction with water (20 mL) in an ultrasonic bath (60 °C, 30 min) purification by adding acetonitrile (20 mL), Carrez I and Carrez II (500 μL each), centrifugation (4,500×g, 10 min), supernatant filtration through a membrane filter
Merck LiChrospher 100 CN column, 250 × 4 mm
Mobile phase, 50% acetonitrile in 1% acetic acid isocratic for 5 min, following rinsing with 100% acetonitrile for 5 min, 0.7 mL/min (spilt 1:5), runtime, 10 min; Inj, 10 μL or 40 μL
LC–MS/MS; LOD, <10_g/kg; LOQ, <30_g/kg;
Electrospray voltage, 5.5 kV; source temperature, 350 °C; m/z transitions (collision energy); AA, 72 > 72, 55, 44 (18 eV); IS, 75 > 75, 58, 44 (18 eV)
Andrzejewski et al. 2004
Spiked with IS, extraction with HPLC grade water, centrifuge tubes capped and shaken/vortexed (30 s), centrifugation, aliquot transfer to a Maxi-Spin PVDFb filtration tube (0.45 μm) and centrifugation Conditioning (Oasis HLB 6 cc cartridge) with methanol and water (3.5 mL each), filtered extract (1.5 mL) loaded, water (1.5 mL) elution and effluent transfer onto the second cartridge (Bond Elut-Accucat), a mark placed on the outside of the tube at a height equivalent to 1 mL of liquid above the sorbent bed, conditioning with methanol and water (2.5 mL each), sample loaded and collected
Synergi Hydro-RP 80A column, 250 mm × 2 mm
Mobile phase, 0.5% methanol in water, 0.2 mL/min; run time, 10 min; Inj, 20 μL
LC–MS/MS; LOD, 10 μg/kg;
Capillary voltage, 4.1 kV; cone voltage, 20 V; source temperature, 120 °C; desolvation temperature, 250 °C; m/z transitions (collision energy); AA, 72 > 72 (5 eV), 55 (10 eV), 27 (19 eV); IS, 75 > 75 (5 eV), 58 (10 eV), 29 (19 eV)
Gökmen et al. 2005
2 g sample + IS + 10 ml methanol, mixture centrifugation (10 min at 11180×g), supernatant clarification with Carrez I and IIa and centrifugation, drying of 2.5 ml supernatant under nitrogen, resolving residue in 1 ml water, Oasis HLB cartridge preconditioning with 1 ml of methanol and 1 ml of water, extract loaded (1 ml) and collected, filtration through a 0.45 μm syringe filter
Mobile phase, 1.0 or 0.5 ml/min; column temperature, 25 °C; Inj, 20 μl
DAD, LOD, 2.0 μg/ml, LOQ, 4.0 μg/kg
226 nm with peak spectra 190–350 nm
dC18, 250 × 4.6 mm
HILIC, 250 × 4.6 mm
Zorbax SIL, 250 × 4.6 mm
bond C18, 250 × 4.6 mm
HiChrom 5 C18, 300 × 4.6 mm
Luna C18, 250 × 4.6 mm
Synergi MAXRP, 250 × 4.6 mm
Zhang, et al. 2007
Various food products
1.5 g sample + IS, 10 min standing, adding 20 ml petroleum ether, 10 min shaking, removal of petroleum ether and repeating defatting, adding 7 ml NaCl (2 mol/L) and 20 min shaking, centrifugation (15 min at 15,000 rpm), extracting the residue with 8 ml NaCl, extracting the NaCl solution with 15 ml ethyl acetate for three times, drying organic phase under nitrogen, adding 1.5 ml water to residue, preconditioning of Oasis HLB cartridge (3.5 ml methanol and 3.5 ml water), loading (1.5 ml) and extracting
Atlantis dC18 column; 210 × 1.5 mm
Mobile phase, 10% methanol /0.1% fomaric acid in water; 0.2 ml/min; column temperature, 25 °C
Capillary voltage, 3.5 kV; cone voltage, 50 V; source temperature, 100 °C; desolvation gas temprature, 350 °C; desolvation gas flow, 400 L/h nitrogen; cone gas flow, 45 L/h nitrogen; argon collision gas pressure, 3 × 10-3 mbar; m/z transitions (collision energy); AA, 72 > 72 (1 eV), 72 > 55 (6 eV), 72 > 44 (9 eV), 72 > 27 (15 eV); IS, 75 > 75 (1 ev), 75 > 58 (6 eV), 75 > 30 (15 eV)
Kim et al. 2007
Rice, bread, corn chips, potato chips, biscuits, candy, coffee
10 g sample + IS+ 98 ml water, 20 min shaking, centrifugation (10 min at 9,000 rpm), conditioning C18 solid-phase extraction cartridge (5 ml methanol and 5 ml water), loading and collecting, filtration through 0.45 μm membrane
Aqua C18 HPLC column, 2 × 250 mm
Mobile phase, aqueous 0.2% acetic acid and 1% methanol; 0.2 ml/min; run time, 14 min; Inj, 20 μl
LC-MS/MS; LOQ, 2 μg/kg
Capillary voltage, 4.2 kV; source temperature, 120 °C; desolvation temperature, 240 °C; desolvation gas flow rate, 650 L/h nitrogen; argon gas pressure, 2.5 mbar; m/z transitions; AA,72 > 55; IS, 75 > 58
Genga et al. 2008
Fried potato chips, biscuits, Chinese fried/baked foods
2 g sample+ 10 ml methanol 75%, treating with Carrez I and II, shaking (45 min t 100 rpm), centrifugation (10 min at 10000 rpm), evaporating 5 ml of supernatant to 1 ml under nitrogen, preconditioning Oasis HLB cartridge (5 ml methanol and 5 ml water), loading and eluting by 2 ml 10% methanol, filtration of extract with 0.45 μm syringe filter
HC-75 H+ column, 305 mm × 7.75 mm
Mobile phase: sulfuric acid (5 mM); 0.6 ml/min; Inj, 20 μl; column temperature, 50 °C
HPLC-DAD; LOD, 30 μg/kg
Liu et al. 2008
1 g sample+ IS + 9 ml water, 20 min shaking, adding 10 ml acetonitrile+4 g anhydrous magnesium sulfate + 0.5 g of sodium chloride, 1 min shaking, centrifugation (5 min at 5,000 rpm), separating of acetonitrile layer and drying under nitrogen, dissolving the residue in 05 ml water, filtering through 0.45 μm syringe filter, preconditioning of Oasis MCX SPE cartridge (2 ml methanol and 2 ml water), loading and collecting, filtration through 0.22 μm syringe filter
ODS-C18 column, 250 mm × 4.6 mm
Mobile phase, 10% acetonitrile and 90% water containing 0.1% formic acid; 0.4 ml/min; Inj,
LC-MS/MS; LOD, 1 ng/ml; LOQ, 5 ng/ml
Capillary voltage, 1 kV; cone voltage, 20 V; 1source temperature, 110 °C; desolvation temperature, 400 °C; desolvation gas flow, 600 L/h nitrogen; cone gas flow, 50 L/h; argon collision gas pressure to 2 × 10-3 mbar; m/z transitions (collision energy)AA, 72 > 55 (13 eV); IS, 75 > 58 (13 eV)
Gökmen et al. 2009
Cookie, potato crisp, bread crisp
Aqueous extraction, 1 g sample + IS + 9 ml of 10 mM fomaric acid, treated with Carrez I and II, centrifugation (10 min at 5,000 rpm), four stage extraction of supernatant without Carrez clarification, preconditioning Oasis MCX cartridge, extract loaded and collected, filtration through 0.45 μm nylon filter.
Atlantis T3 column, 150 mm × 4.6 mm
Mobile phase, 10 mM fomaric acid; 0.3 ml/min; column temperature, 25 °C
LC-MS, LOQ, 15 μg/kg
Capillary voltage, 2 kV; corona current, 5 μA; drying gas temperature, 350 °C; m/z transitions (collision energy); IS, 72
Methanol extraction, 1 g sample+IS+9 ml methanol, centrifugation (10 min at 5,000×g), supernatant treated with Carrez I and II, centrifugation (10 min at 5,000 rpm), four stage extraction of supernatant, drying under nitrogen, reconstitution of residue in 1 ml water, elution through preconditioned Oasis MCX cartridge, filtration through 0.45 μm nylon filter.
Gas chromatography methods
Tareke et al. 2000
10 g sample+ 100 ml water, filtration of extract through glass-fiber filter, purification on carbograph 4 column, addition of standard, bromination
HP PAS 1701 column, 25 m × 0.32 mm
65 °C held for 1 min, ramped at 15 °C/min to 250 °C, held for 10 min; Inj, 2 μL; splitless
GC–MS; LOD, 5 μg/kg
m/z transitions; AA, 152, 150, 108,106; IS, 180, 178
Tareke et al. 2002
Protein-rich and carbohydrate-rich foods
10 g sample+ 100 ml water, filtration of extract through glass-fiber filter, purification on carbograph 4 column, addition of standard, bromination
N,N-dimethyl acrylamide or [13C1]acrylamide
BPX-10 column, 30 m × 0.25 mm
65 °C held for 1 min, ramped at 15 °C/min to 250 °C, held for 10 min; Inj, 2 μL; splitless
GC–MS; LOD, 5 μg/kg
m/z transitions; AA, 152, 150, 106; IS, 180, 155
Ono et al. 2003
Various food products
50 g sample+IS+300-400 ml water, homogenization, centrifugation (20 min at 48,000×g), freezing and melting of supernatant, centrifugation (10 min at 21,700×g), fractionation of 0.5-2 ml supernatant on SPE cartridge, fraction collected and centrifugation (10 min at 27,000×g), filtration of supernatant through 0.22 μm syring filters, centrifugation of filtration with cut-off of 3,000 Da (50 min at 14,000×g)
CP-Sil 24 CB Lowbleed/MS column, 30 m × 0.25 mm
85 °C held for 1 min, ramped at 25 °C/min to 175 °C, held for 6 min, ramped at 40 °C/min to 250 °C, held for 7.52 min
GC–MS; LOD, 0.2 ng/mL
m/z transitions; AA, 52, 150; IS, 155, 153
Hoenicke et al. 2004
Various food products
Weighed into a filter, placed on a Witt’scher pot, equipped with a vacuum pump, defatted by adding iso-hexane (80 mL), spiked with IS, incubation (30 min), extraction with water (20 mL) in an ultrasonic bath (60 °C, 30 min) Purification by adding acetonitrile (20 mL), Carrez I and Carrez II (500 μL each), centrifugation (4500 × g, 10 min), supernatant filtration through a membrane filter
DB-WAX capillary column,30 m × 0.25 mm
70 °C held for 1 min, ramped at 20 °C/min to 230 °C, held for 10 min; Inj, 1 μL, splitless
GC–MS/MS; LOQ, 30 μg/kg
m/z transitions;AA, 89 > 72, 55; IS, 92 > 75
Hamlet et al. 2004
5 g sample+IS+deionised water, 1 min shaking, adding 0.3 ml glacial acetic acid, treating with Carrez I and II, centrifugation (20 min at 1942 g)
Rtx®-50 column, 30 m × 250 μm
65 °C held for 2 min, ramped at 15 °C/min to 250 °C, held for 5 min; Inj, 1 μl; splitless
GC-MS/MS; LOD; 0.01 ng/ml
Ionization mode, negative; argon collision gas pressure, 1.5 mTorr; m/z transitions (collision energy); AA, 149 > 70 (10 eV), 151 > 70 (10 eV); IS, 152 > 73 (10 ev), 154 > 73 (10 eV)
Pittet et al. 2004
Weighted (15 g) into a 250 mL centrifuge bottle, spiked with IS, sample suspended in water (150 mL) and homogenized (30 s), suspension acidified to pH 4–5 by addition of glacial acetic acid (∼1 mL), treated successively with Carrez I and IIc (2 mL each), centrifugation (16,000 g, 15 min), bromination, extract transferred onto a glass chromatography column containing calcinated sodium sulfate and activated Florisil (5 g each), using small aliquots taken from hexane (50 mL), acrylamide derivative eluted with acetone (150 mL), evaporated to ∼2 mL and then to dryness (N2), re-dissolved in EtAcd (400μL), triethylamine added (40μL)filtered through a 0.2 μm microfilter
ZB-WAX capillary column, 30 m × 0.25 mm
65 °C held for 1 min, ramped at 15 °C/min to 170 °C, 5 °C/min to 200 °C, 40 °C/min to 250 °C, held for 15 min; Inj, 2 μL, splitless
GC–MS; LOD, 2 μg/kg;
m/z transitions; AA, 149, 70; IS, 154, 110
Dunovska et al. 2006
Potato crisps, breakfast cereals, crisp bread
3 g sample+ IS+ 4.5 ml deionized water, 30 min ultrasonic bath, adding 24 ml n-propanol, centrifugation (5 min at 11000 g), adding 5 drop olive oil, drying, dissolving the residue in 2 ml MeCN, defatting with 10 and 5 ml n-hexane, mixing 1 ml MeCN with 60 mg PSA sorbent, centrifugation (1 min at 11000 rpm),
INNOWx capillary column, 30 m × 0.25 mm
70 °C for 1.0 min, 20ºCmin−1 to 240 °C (held for 10.5 min); carrier gas, helium, 1.0 ml/min; Inj, 1 μl; pulsed splitless 1.0 min, 4 ml/ min−1
GC–HRTOF MSb; LOQ, 15 and 40 μg kg−1
Acquisition rate, 2 Hz; pusher interval, 33 μs (30303 raw spectra s−1); inhibit push value, 14; time-to-digital converter (TDC), 3.6 GHz; mass range, m/z 45–500; ion source temperature, 220 °C; transfer line temperature, 240 °C; detector voltage, 2,600 V.
Serpen and Gokmen 2007
2 g sample+IS+ 20 ml methanol, Carrez clarification, centrifugation (10 min at 10000 rpm) using 0.45 μm microspin PVDF centrifuge filter
HP INNOWAX column, 30 m × 250 μm
Isothermal for 0 min, ramped at 10 °C/min from 80 to 280 °C, Isothermal for 13 min; flow of carrier gas 1.0 ml/min; Inj, 1 μl; splitless
GC-MS; LOD, 15 ng/g; LOQ; 50 ng/g
Electron ionizing, 70 eV, m/z transitions; AA, 71 > 71,55,27; IS, 74, 58
Zhang et al. 2007
Various food products
1.5 g sample+ IS, 10 min standing, adding 20 ml petroleum ether, 10 min shaking, removal of petroleum ether and repeating defatting, adding 7 ml NaCl (2 mol/L) and 20 min shaking, centrifugation (15 min at 15000 rpm), extracting the residue with 8 ml NaCl
HP INNOWAx capillary column, 30 m × 0.32 mm
110 °C for 1 min, 10 °C/min to 140 °C, 140 °C held for 15 min, ramped at 30 °C/min to 240 °C, and finally isothermal at 240 °C for 7 min; carrier gas, nitrogen; Inj, 1 μL,
GC-MECD; LOD, 10 μg/kg
Ionization mode, positive; m/z transitions; AA, 70, 149, 151; IS, 110, 154
Lee et al. 2007
French fries, potato chips
10 g sample+100 ml water, centrifugation (10 min at 5000 rpm), dilution of 1.5 ml aliquot to 15 ml with water, mixing with 15 ml buffer (pH 7), immersing of SPME fiber
DB-WAX silica capillary column, 30 m × 0.25 mm
Ramped at 15 °C/min from 80 °C to 220 °C, held at 220 °C for 2 min; carrier gas, helium, 1 mL/ min; elution time, 9.88
GC-PCI-MS/MS; LOD, 0.1 μg/L
Mass-to-charge ratio (m/z) scan range from 40 to 100 u
Extraction, Purification, and Derivatization
The high water-solubility of acrylamide means that extraction from foods using plain water is very effective, with no need for pH adjustment (Eriksson 2005).
Water at room temperature has been used as an extractant in most LC methods (Becalski et al. 2003; Rosén and Hellenäs 2002; Tareke et al. 2000). Heating or sonicating during the extraction should be avoided as this may generate the SPE columns used in further clean-up steps (US Food and Drug Administration 2003). Nevertheless, Ahn et al. (2002) used heated water (at 80 °C) without any problem during clean-up.
Purification of water extracts for acrylamide analysis is based the following principles: purification with SPE columns and chemical purification (deproteination) (Eriksson 2005). However, the extraction solvent could be a mixture of water and organic solvents such as n-propanol or 2-butanone (Biedermann et al. 2002b). Recovery rates of 68–75.4% have been reported when pure methanol was applied for the extraction of baked food products (Tateo and Bononi 2003). Also, the extracted amount of acrylamide, using Soxhlet extraction with methanol, far exceeded that by other extraction techniques (>90%). However, one drawback is the long extraction time of 10 days while no information is available on the potential acrylamide formation during the extraction (Wenzl et al. 2003). To overcome these problems, some researchers have extracted acrylamide selectively by using ASE without co-extracting the starch, which leads to the simultaneous simplification of the subsequent sample clean-up (Brandl et al. 2002). Although solubility of acrylamide in most organic solvents is lower than water, dichloromethane appears to be a promising extraction solvent, particularly addition of 2% ethanol as a modifier improves the results considerably. Re-extraction of the organic phase with water delivers an aqueous solution which is almost free of any interfering matrix components. For instance, fatty components remain in the organic phase; thus, defatting of the sample prior to extraction is not necessary any more. Applying this method, nearly all of the target foodstuffs could be analyzed in a similar fashion with satisfactory results (Brandl et al. 2002; Dionex 2004).
In using the ASE method, pure water, 10 mM formic acid solution, and acetonitrile were tested as the extraction solvent in LC methods of acrylamide determination. Pure water extracts showed lower recoveries than the formic acid, but the formic acid extracts had a lower stability. Acetonitrile extracts were cleaner, as less material was co-extracted from the sample matrix. With three extraction cycles of 4-min durations, a yield of 95% in the first extract and an additional 8% in the second extraction of the same sample using 10 mM formic acid were achieved (Dionex 2004). However, a mixture of water and acetone has also been reportedly used as the extractant (Takatsuki et al. 2003; Fauhl et al. 2002). Other researchers have also used the ASE device (Cavalli et al. 2002; Höfler et al. 2002).
Different mechanical methods can be used for the initial extraction steps that include shaking at high speeds on a horizontal shaker (Becalski et al. 2003), using a rotating shaker (US Food and Drug Administration 2003), occasional swirling (Ahn et al. 2002; Takatsuki et al. 2003), and mixing with a blender or mixing on a vortex (Fauhl et al. 2002).
After extraction, the aqueous phase is centrifuged and different laboratories have reported different centrifugation conditions, as described in Wenzl et al. (2003). Becalski et al. (2003) and FDA (2003) recommended combined centrifugation and filtration using a 5 kDa cut-off Centricon Plus-20 and 0.45 μm PVDF filters, respectively. Ono et al. (2003) used centrifugation filters with a 3 kDa cut-off after clean-up of the extract by SPE.
To control the recoveries achieved and to keep track of possible losses during the extraction and purification steps, an internal standard is added to the food-extraction mixture. Similar to the GC-MS methods, isotopically labeled [13C3]-acrylamide (Tareke et al. 2002; Becalski et al. 2003), [d3]-acrylamide (Becalski et al. 2003; Ahn et al. 2002; Rosén and Hellenäs 2002; Ono et al. 2003), and [13C1]-acrylamide (Takatsuki et al. 2003) have been used.
Most purification procedures consist in combining several SPE. For instance, a combination of three different cartridges: mixed-mode anion exchange, mixed-mode cation exchange, and graphitized carbon have been used (Becalski et al. 2003). Takatsuki et al. (2003) also used a similar combination of SPE cartridges for the clean-up of samples, which were measured by LC-MS with column switching. Also, a combination of SPE and filtration and/or ultracentrifugation has been used to avoid blockage of the chromatographic system (Wenzl et al. 2003). However, Höfler et al. (2002) reported that both SPE and liquid–liquid extraction did not lead to any significant improvement in the analysis. Therefore, filtration through a 0.22-μm nylon filter is the only sample treatment used as clean-up procedure after extraction and before applying to HPLC. In contrast, other laboratories added acetonitrile to the aqueous extract and pipetted 0.5 ml Carrez I and Carrez II onto the sample in order to obtain a clear sample (Wenzl et al. 2003).
One special aspect of the extraction procedure involves the swelling of the matrix in order to provide better access for the extraction solvents to potentially adsorbed or enclosed acrylamide. However, the side-effect associated with swelling is that it provides some time for the development of matrix/internal standard interactions. For this reason, the homogenized sample is mixed with water and an internal standard solution and kept at a pre-specified temperature for 10–20 min. Depending on the matrix, swelling yielded an increase in analyte recovery of up to 100-fold (Biedermann et al. 2002b).
Although extraction at room temperature provides satisfactory results, hot water (60–80 °C) can be used to enhance the extraction. Increased recovery has also been observed by treating the sample in an ultrasonic bath (30 min at 60 °C; Schaller 2003). Problems with the high viscosity of the sample/water extraction mixture have been reported to be solved by the addition of small amounts of amylase to the mixture.
In derivatization methods, acrylamide is converted to 2,3-dibromopropionamide which is volatile and can be detected on a GC with an electron capture or an alkali flame detector (Tekel et al. 1989; United States Environmental Protection Agency 1996). This bromination is mostly performed by ionic reaction (Hashimoto 1976; Arikawa and Shiga 1980). The analysis has been suggested to be performed on the more stable 2-bromopropenamide obtained after debromination of 2,3-dibromopropionamide (Andrawes et al. 1987; Martin et al. 1990). Applications of SPE columns to obtain concentrated samples or utilizing a more sensitive derivatization technique may increase the possibility for determination of residual acrylamide in many types of food (Kawata et al. 2001; Pérez and Osterman-Golkar 2003).
Bromination of acrylamide has the advantage that a more volatile compound is produced and the selectivity of determination is enhanced (Wenzl et al. 2003). However, some derivatization approaches are laborious and time-consuming. The procedure first reported by Hashimoto (1976) is carried out by adding a pre-prepared bromination solution containing potassium bromide, hydrogen bromide, and bromine to either the pre-treated or raw aqueous extracts (Tareke et al. 2002; Ahn et al. 2002; Castle 1993; Castle et al. 1991; Ono et al. 2003). In this method, the yield of 2,3-DBPA is constant and >80% when the reaction time is more than 1 h (United States Environmental Protection Agency 1996).
Nemoto et al. (2002) improved the method by using different derivatization reagents including potassium bromide and sodium bromate in an acidic medium. The necessity of additional sample pre-treatment depends upon the matrix. Matrices such as carbohydrate rich foods (e.g., potato crisps or bread) require additional fractionation steps (Tareke et al. 2002; Tareke et al. 2000). Usually, the raw extract is subjected to fractionation on a graphitized carbon black cartridge.
Bromination is frequently carried out overnight at 0 °C or slightly above the freezing point of water. It has also been stated that application of isotopically labeled internal standards allowed a reduction in the reaction time from overnight to 1 h. This is in accordance with the methods proposed by other scientists (Ono et al. 2003; Nemoto et al. 2002).
The excess of bromine is removed after the reaction by titration with sodium thiosulfate solution (0.7–1 M) until the solution becomes colorless. The brominated acrylamide is less polar compared with the original compound and, therefore, non-polar organic solvents (usually ethyl acetate or a mixture of ethyl acetate and cyclohexane) are used for the extraction of the analyte from the aqueous phase. Gertz and Klostermann (2002) reported that, on a DB-5 MS column, a transformation of 2,3-DBPA to 2-BPA does not take place, so that it is not necessary to transform the dibrominated compound into the more stable 2-monobromopropenamide by adding triethylamine. However, more recent reports confirm that, in acrylamide analysis with bromination and detection by GC-ECD in HP-INNOWAX capillary column, 2-BPA rather than 2,3-DBPA was chosen as the quantitative analyte because the peak response of former was nearly 20 times higher than that of the latter (Zhang et al. 2006).
Defatting has to be included in some sample preparations because of the influence of high fat content on the analysis. In one method, the fatty compounds were removed by extracting with hexane or by using graphitized carbon cartridges after swelling (Tareke et al. 2000; Tareke et al. 2002). Other methods include a phase separation by centrifugation followed by removal of the water fraction by azeotropic distillation (Biedermann et al. 2002b; Tateo and Bononi 2003).
The effects of other factors, such as pH, on acrylamide extraction have been studied. Changes in pH have been found to have a significant effect on acrylamide extraction efficiency. A higher amount of acrylamide (three to four times) was observed in food samples by changing pH towards the alkaline pH (pH>12). One reason might be that, during normal water extraction, polyacrylamide sterically hinders all of the acrylamide from getting into the solution. Alkaline pH can change the structure of the matrix and facilitate the free acrylamide to get into the solution. Also, increasing pH releases the chemically bound acrylamide (bound with protein and carbohydrate) to become available for analysis with this extraction technique (Kim et al. 2001; Svensson et al. 2003).
Phase separation is usually carried out by centrifugation of the sample. Further clean-up is also performed by fractionation of the organic extraction on silica–gel cartridges (Castle 1993; Castle et al. 1991). Since silica is of high water adsorptivity, ethyl acetate has to be dried or replaced by cyclohexane to avoid any change in silica activity. Florisil is also used as an adsorbent and a mixture of acetone and hexane as the elution solvent (Nemoto et al. 2002). Alternatively, gel permeation chromatography on Bio-Beads S-X-3 gel is performed as the final sample clean-up (Tareke et al. 2000). Recently, drying of the extract has been carried out by adding sodium sulfate (Tareke et al. 2002; Ahn et al. 2002). In addition, removal of residual water eliminates or decreases the effect of interferences from water-soluble co-extractants. The solvent volume has to be reduced to 30–200 μl to reach limits of detection in the range of 1–5 μg kg-1 before injection into the GC column (Tareke et al. 2002; Tareke et al. 2000; Castle 1993; Castle et al. 1991).
Analyte Separation, Detection, and Quantification
Owing to the higher polarity of the non-derivatized acrylamide, columns with polar phase, e.g., polyethylenglycol, have to be used. A total of 1–2 μl samples is usually injected in splitless mode, and analyte separation is performed on standard GC capillary columns with a length of 30 m and an internal diameter of 0.25 mm (standard in GC-MS). The initial oven temperature is normally adjusted to 60 °C to 85 °C, and the heating rate is 15 °C min-1. The final oven temperature is usually about 250 °C. This is while columns with middle to high polarity can be used in the case of derivatized acrylamide. However, it is possible to inject sample extracts into the GC in the split mode (Biedermann et al. 2002b; Tateo and Bononi 2003; Wiertz-Eggert-Jörissen (WEJ) GmbH 2003).
The major drawback of GC analysis without derivatization is the lack of characteristic ions in the mass spectrum of underivatized acrylamide. In the electron ionization mode, the major fragment ions are at m/z 71 and 55, respectively, which are also used for quantification. Co-extracted substances such as maltol or heptanoic acid produce almost the same fragmentation pattern and may, therefore, interfere. Selectivity can be increased by chemical ionization using methane as the reagent gas (Biedermann et al. 2002b).
The original method using GC-MS with bromination is based on adding methacrylamide as an internal standard to the homogenized sample and then producing the derivative 2,3-dibromo-2-methylpropionamide (Castle et al. 1991). Alternatively, methacrylamide can be derivatized separately and added to the sample directly before the final adjustment of the solvent volume (Castle 1993). The latter is preferred because methacrylamide acts as a chromatographic internal standard, which means that it can monitor potential changes in the performance of the instrument. However, Castle (2003) reported a large difference between the reaction kinetics of the bromination reactions of acrylamide and methacrylamide.
Quantification is performed by adding different kinds of internal standards, ranging from propionamide to isotropically labeled acrylamide. LOD is reported to be about 10–50 mg kg-1. Acrylamide may also be determined by positive chemical ionization with ammonia as the reagent gas and tandem mass spectrometric detection of the daughter ions released from the single charged molecular ion adduct. Thus, LOD can be reduced by GC-MS/MS to 1–2 μg kg-1.
The methods usually used for acrylamide quantification include external and internal standard methods. The external quantitative analysis, however, revealed poor reproducibility. Zhang et al. (2006) suggested an improvement of the quantitative method by introducing the internal standard, which is the most commonly used quantification method for acrylamide. The commonly used internal standards include isotope-labeled internal standard (e.g., 13C3-acrylamide or 2H3-acrylamide, etc.) and non-isotope-labeled internal standards (e.g., methacrylamide, etc.). Although the isotope-labeled internal standards are the most ideal ones of the type, they can only be used in MS-based analysis and satisfactory repeatability cannot be achieved until isotope-labeled acrylamide is used. This could be due to the differing stability of the compounds and incomplete derivatization of structurally different internal standards. Due to the large differences reported between the reaction kinetics of the bromination reactions of acrylamide and methacrylamide, a long bromination reaction time is required when methacrylamide is used as the internal standard (Wenzl et al. 2003; Castle and Eriksson 2005). The method of standard addition is also an accurate quantification method and especially useful when the matrix of the sample is very complex and extraction yields strongly vary (Basilicata et al. 2005; Ito and Tsukada 2001). In cases where isotopically labeled acrylamide is used, repeatability of results is achieved by adding N,N-dimethylacrylamide. The properties of N,N-dimethylacrylamide are obviously far different from those of acrylamide. Consequently, the coefficient of variation (CV) of the acrylamide recovery of spiked samples decreased from 26% to 7.5% when applying [13C3]-acrylamide (Tareke et al. 2002). Castle (1993) reported that 2,3-dibromopropionamide might eliminate hydrogen bromide during injection or chromatographic separation. Others used dehydrobromination instead of 2,3-dibromopropionamide (Andrawes et al. 1987; Nemoto et al. 2002; Takata and Okamoto 1991). Another reason for the large CVs might be incomplete derivatization of structurally different internal standards (methacrylamide and N,N-dimethylacrylamide). Meanwhile, most laboratories use [D3]-acrylamide, [13C3]-acrylamide or both together as internal standards.
The quantification is usually performed by the method of standard addition as follows: a set of GC peak areas of the analyte obtained for each sample (one for unspiked and three for spiked portions with different levels of standard solutions) are plotted as along the y-axis, while the quantities of standard substances in the portions are plotted as the x-axis. A calibration curve is then prepared using the linear regression method to calculate the amount of the analyte in the spiked portion of the sample (Zhu et al. 2008).
For the chromatographic separation of acrylamide, most scientists have used reversed-phase chromatography (different C18 columns; European Union Risk Assessment Report 2002; Rosén and Hellenäs 2002; Takatsuki et al. 2003). Different reversed-phase columns have been compared. Primisphere C18-HC is recommended because it provides sufficient retention time for acrylamide to minimize matrix interference. An alternative to reversed-phase columns is the IC. An IonPac column is a combined ion exchange with size exclusion chromatography. The advantage is that there is a good separation of acrylamide from matrix compounds of even untreated sample extracts (Ahn et al. 2002; Cavalli et al. 2002; Höfler et al. 2002).
For detection of acrylamide after LC separation, tandem mass spectrometry is most often the method of choice. There are just a few exceptions in which UV at 202 nm (incurring a lack of selectivity) and single quadruple MS (in the single ion monitoring mode) are used (Höfler et al. 2002). This lack of selectivity would hamper the determination of acrylamide in complex matrices. A solution could be the use of two-dimensional (Fauhl 2003) and/or multidimensional LC (Takatsuki et al. 2003) that use four different columns for separation of the analyte from the interference.
LC-MS/MS, working in MRMs, in which the transition from a precursor ion to a product ion is monitored, has a high selectivity. MRM means that the transition from a precursor ion, which is separated in the first quadruple, to a product ion, generated by collision with argon in the second quadruple, is monitored in the third quadruple. The transition 72 → 55 has been usually selected for quantifying acrylamide because it shows a relatively high intensity (Becalski et al. 2003; Ahn et al. 2002; Rosén and Hellenäs 2002; Tareke et al. 2000). Other transitions, such as 72 → 54, 72 → 44, and 72 → 27, have been used in some cases for configuration. For the detection of the isotopically labeled acrylamide used as internal standard, the monitored transitions are 75 → 58 for [D3]- and [13C3]-acrylamide and 37 → 56 for [13C1]-acrylamide. Despite the selectivity offered by MS/MS, interference can still occur. Peaks showing identical retention times to acrylamide and deuterated acrylamide have been observed. This problem could be solved by increasing the pH of the solution from which acrylamide was extracted into an organic solvent (e.g., the ASE device used during extraction; Swiss Federal Office of Public Health 2002). Becalski et al. (2003) also reported the existence of an early eluting compound that interferes when transition 72 → 55 is used for the detection of acrylamide. Increasing the column length from 100 to 150 mm and applying Isolute Multi-Mode cartridges during sample preparation eliminated the problem. Further details on the chromatographic conditions and the optimum parameters used for the MS/MS and UV detectors are reported in Wenzl et al. (2003).
Strategies for Reduction of Acrylamide Levels in Food
Different strategies for reducing acrylamide levels in food have been suggested. Removing or reducing of reactants is one of them. When one of the reactants (asparagine or glucose) is at lower concentration, formation of acrylamide will be reduced. Decreases in asparagine content may be achieved by: (a) selecting cultivars (e.g., potatoes, cereal grain) that contain lower levels of asparagine; (b) elimination of enzymes which control biosynthesis of asparagine by suppressing genes that encode them; (c) hydrolysis of asparagine to aspartic acid and ammonia by acid- and/or asparaginase/amidase-catalysis; (d) modification of asparagine to N-acetylasparagine via acetylation, so formation of N-glycoside intermediates, which form acrylamide, is prevented (Friedman 1978, 2001).
Meanwhile, researchers have determined two possible ways to reduce the level of sugar in potatoes. During storage of potatoes, the amount of sugar increases, thus using fresh potatoes could result in less acrylamide being formed. Also, storing potatoes below 8–10 °C can increase the formation of reducing sugars, and the presence of these reducing sugars together with asparagine may lead to acrylamide formation. The variety of potato with relatively lower amounts of reducing sugars and asparagine also affects the amount of acrylamide formation (FAO/WHO 2004).
In another procedure, the formation of acrylamide will be reduced with disruption of reaction. There is a time–temperature relationship to the formation of acrylamide in food, thus changing the temperature or duration of cooking will affect the level of acrylamide. It has been suggested that, when the temperature of food rises above 120 °C, the rate of acrylamide formation increases rapidly with temperature over a limited range (Claus et al. 2008). Acrylamide formation in food is also pH-dependent and optimum pH for acrylamide formation in food is about 7. In acidic pH, acrylamide formation is inhibited. Lowering the pH of the food system to reduce acrylamide generation may attribute to protonating the α-amino group of asparagine, which subsequently cannot engage in nucleophilic addition reactions with carbonyl source (Zhang and Zhang 2007). Other inhibitors of acrylamide formation are the asparaginase enzymes that disrupt the formation of acrylamide. Another factor is water activity that seems to be a critical factor (Taeyman et al. 2004).
Destroying and/or trapping of acrylamide after its formation is another strategy which can be done via (a) hydrolysis of the amide group of acrylamide to acrylic acid and ammonia by acid- or enzyme-catalysis; (b) polymerization of monomers of acrylamide to polyacrylamide in processed foods by means of using UV light or radiation (Friedman 1997); (c) reaction of acrylamide with SH-containing amino acids, esters, peptides, and proteins (Friedman 1996). Also, some compounds like NaCl and CaCl2 could decrease the amount of acrylamide formed (Açar et al. 2010; Pedreschi et al. 2010).
Only a limited number of methods have been so far proposed in the literature on the determination of acrylamide in food products. As for the recognition of methods, mainly two general methods of analysis (LC-MS/MS or GC-MS) are used, and it is still difficult to determine which one is more reliable.
By comparing methodologies, large differences are found among the extraction procedures and clean-up strategies, e.g., variation in extraction, in swelling conditions, in temperature and time of extraction, in mechanical treatment, and in centrifugation or the use of SPE for both GC- and LC-based methods, especially in LC sample preparation. GC-MS after bromination is the best approach so far, because this method is a relatively mature coupled technique with adequate sensitivity and multiple ion conformation.
Application of GC-MS/MS or coupling to a high-resolution MS would even further lower the detection limit of certain foods to 1–2 μg/kg. Determination of acrylamide using LC-MS/MS may avoid derivatization and has the advantages of rather high sensitivity and stability.
Also, research should be focused on cheap, convenient, and rapid screening methods that are reliable and robust which could be employed in most laboratories. In this respect, GC-ECD method has been developed for identification and quantification of derivatized acrylamide in heat-processed starchy foods, while it requires a relatively low-cost instrumentation to perform compared with MS detection-based methods. The ASE method provides a fast and efficient extraction of acrylamide from various food samples along with simple and rapid sample preparation without SPE clean-up and concentration prior to GC-ECD analysis. In addition, the standard addition method has been reported as a suitable quantification method for the determination of acrylamide in heat-processed foods.