Factors Affecting the Formation of Acrylamide in Coffee

Vol. 22, Special Issue Czech J. Food Sci.

K. BAGDONAITE* and M. MURKOVIC

Graz University of Technology, Department of Food Chemistry and Technology, Graz, Austria, E-mail: [email protected]

Abstract: Four types of green coffee beans (Robusta and Arabica) were roasted in a laboratory roaster and in an oven. The samples were analysed for acrylamide using liquid chromatography with UV detection. Significant difference in acrylamide content was observed in different coffee types. Robusta coffee beans, roasted to different degrees of browning contained more acrylamide than Arabica varieties. Roasting time and temperature had a great influence on the acrylamide formation in coffee beans. Coffee beans roasted for longer time had less acrylamide. Additionally, coffee beans roasted at higher temperatures contained less acrylamide compared to those roasted at lower temperatures.

Keywords: acrylamide; coﬀee; roasting; LC-UV

INTRODUCTION fee beans. Some studied proved that AA is formed via the Maillard reactions [7]. It is a result of the Coffee is known worldwide for its flavour and reactions between free amino acids and reducing harmonic taste. A long way of preparation of coffee sugars [8]. The main precursors of AA seem to be beans since picking them from the coffee trees till asparagin and reducing sugars (glucose, fructose roasting has an influence on the cupping quality and etc.) [8, 9]. and taste. What kind of coffee beans to choose The aim of this study was to investigate the influ- and degree of roasting depends on consumer’s ence of coffee type, roasting conditions (time and taste and traditions. Many chemical reactions take temperature) and AA formation in coffee beans. part when coffee beans are roasted, the typical flavour and taste of coffee forms during this proc- EXPERIMENTAL ess, and the physical and chemical properties of the beans change. During roasting the volume Chemicals and solvents. n-Hexane (HPLC grade) of a bean increases, moisture content decreases. and acetonitrile (HPLC grade) were obtained from The aroma and taste compounds in coffee beans LGC Promochem (Germany), formic acid 98–100% form when Maillard reactions (or non-enzymatic purity from Riedel-de Haen (Seelze, Germany). browning reactions), Streckers degradation, sugar Water was obtained from purification system Sim- degradation, minor lipid degradation reactions and plicity 185 (Millipore, Billerica, MA, USA). Stand- interaction between intermediate decomposition ard of acrylamide 99+% purity was obtained from products take place [1]. Recent studies showed MERCK-Schuchardt (Hohenbrunn, Germany). that not only desirable compounds can be found Green coffee beans. Green coffee beans of Ara- in roasted coffee. Acrylamide (AA), a neurotoxic bica (Coffea arabica) and Robusta (Coffea canephora substance to both rodents and humans and probable Pierre) species were tested: Camerun Robusta carcinogen to humans [2] was detected in fried, (C. canephora, dry-processed), Santos Brasil, Co- deep-fried and baked foods [3, 4]. AA was also lumbian Excelso and Uganda Organico Biocoffee detected in commercially available coffee powder, (C. arabica, wet-processed). instant coffee and brewed coffee [5]. Some stud- Roasting of coffee. The green coffee beans were ies showed that fresh or boiled foods contain no either heated in a thermostated oven at 220–260°C quantifiable levels of AA [6], it seems that roasting for 5–15 min, or roasted in a small scale coffee is the only probable way for AA formation in cof- roasting maschine (Wheeling, IL, USA).

22 Proc. Chemical Reaction in Food V, Prague, 29. 9.–1. 10. 2004 Czech J. Food Sci. Vol. 22, Special Issue

Typical sample preparation and extraction for The identification of AA was done by comparing analysis of AA. 80 g of each type of green coffee the retention time with the standard substance. beans were roasted in a coffee roaster. After roast- Additionally, for identification of AA LC-MS was ing and cooling, coffee was ground immediately used [10]. in a coffee grinder. 5.00 ± 0.01 g of coffee were taken to the 250 ml flask and fat was removed by RESULTS AND DISCUSSION extraction with 25 ml of n-hexane for 10 min. The procedure was performed in duplicate. The resi- Coffee beans roasted in a roaster. Green coffee dues of n-hexane were removed under a stream beans were roasted in a laboratory roaster accord- of nitrogen. ing to four roasting programmes. During roasting For extraction of acrylamide 10 ml of water were the temperature of all roasting programmes did added to 0.5000 ± 0.0001 g of dried coffee in a cen- not exceed 250°C. However, the time of roasting trifuge tube. After the extraction for 30 s samples differed: 300 s (programme 4), 450 s (programme 6), were centrifuged at 4000 U/min for 30 min and 720 s (programme 8) and 870 s (programme 10). filtered through a filter. Then the supernatant was Cooling always took 300 s. After roasting samples purified by SPE. The SPE cartridge was conditioned were prepared fo HPLC analysis and the content of with 3 ml methanol and 3 ml water. As acrylamide AA was measured. Data are shown in Figure 1. is not retained the first 2 ml of the eluting solution Robusta (Camerun) was shown to have the high- were discarded and the rest collected. Until HPLC est amount of AA formed during roasting. Arabica analysis the samples were stored at 4°C . coffee beans had lower amounts of AA compared Concentration of acrylamide in each sample to Robusta coffee beans. High quality Arabica was calculated from a calibration curve. The limit coffee beans (Columbian Excelso and Uganda of detection of 5 ng/ml was calculated using the Organico Biocoffee) had a lower amount of AA software ValiData V1.01. compared to lower quality Arabica coffee beans HPLC analysis of acrylamide. For HPLC anal- (Santos Brasil). ysis 10 µl of the purified extract were injected Heated coffee beans in an oven. Green coffee into an Agilent 1100 liquid chromatograph with a beans of Robusta (Camerun) were heated on a UV-detector, equipped with an Ion Pac ICE AS-1 glass dish in an oven at 220, 240 and 260°C for 5, (Dionex, Sunnyvale, CA, USA). The column was 10 and 15 min. Before heating the glass dish was eluted isocratically with a flow of 0.400 ml/min preheated for 10 min. After heating the dish with using a mobile phase of 30% acetonitrile in 3mM a sample was immediately put on ice for 15 min. aqueous formic acid. The effluent was monitored Figure 2 shows the AA content in coffee beans for its absorption at 202 nm. roasted in the oven.

4000

3500 ) g /

g 3000 n ( d

e 2500 m r o

f 2000 e d i 1500 m a l y r 1000 c

A Camerun Robusta 500 Santos Brasil 0 Columbian Excelso Figure 1. Acrylamide content (ng/g) in 4 Uganda Organico Biocoffee 6 8 coﬀee beans roasted according to four 10 Programme roasting programmes

Proc. Chemical Reaction in Food V, Prague, 29. 9.–1. 10. 2004 23 Vol. 22, Special Issue Czech J. Food Sci.

Figure 2. AA content in Robusta coﬀee beans heated in an oven at 220–260°C for 5–15 min 500 ) g / 400 g n ( d e 300 m r o f e d

i 200 m a l y r

c 100

A 5 10 0 Time (min) 15 220 240 260 Temperature (°C)

The highest concentration is detected in beans provided as is and no guarantee or warranty is given that roasted at 240°C for 5 min. Beans roasted at 220 the information is fit for any particular purpose. The user and 260°C for 5 min contained the highest amounts thereof uses the information at its sole risk and liability. of AA. Roasting for longer times led to a reduc- tion of the AA concentration. AA seems to form in References highest concentrations during the first five minutes of roasting, but the amount of AA decreases with [1] I�� A., V�� R. (1995): Espresso Coffee: The Che- increasing the time of roasting. By using higher mistry and Quality. London. temperatures and a shorter time of roasting lower [2] Overall Evaluations of Carcinogenicy to Humans concentrations of AA were formed. h�p://193.51.164.11/monoeval/crthgr02a.html (checked on 16 03 2004). CONCLUSIONS [3] T�� E., R�� P., K�� P., E�� S., T�� M. (2002): J. Agric. Food Chem., 50: Our experiments showed a significant difference 4998. of AA in roasted coffee beans. Time of roasting and [4] R�� J.A.G., A�� D., G�� M.L., N�� high temperatures had an influence on the forma- D., M�� S.M. (2003): J. Agric. Food Chem., 51: tion of AA. AA concentration varied depending on 7547. roasting time/temperature value. The AA content [5] A�� D., R�� J.A.G., G�� M.L., M�� varies depending on coffee type. This is in strong S.M. (2004): J. Agric. Food Chem., 52: 1996. agreement with other studies [5]. [6] S�� K., A�� L., B�� W., G�� Green coffee beans after processing contain usu- A., H�� K.-E., L�� Y., R�� J. (2003): Food ally not more than 12% of moisture [1]. The precur- Chem. Toxicol., 41: 1581. sors – free amino acids and reducing sugars – vary [7] Y�� V.A., W�� A., L�� C.P. (2003): depending on coffee spieces [1, 11], thus the AA J. Agric. Food Chem., 51: 1753. concentration after roasting can differ. Reactions [8] B�� A., L�� B.P.-Y., L�� D., S�� S.W. which can occur during roasting are very complex (2003): J. Agric. Food Chem., 51: 802. and more research is required to investigate the [9] Z�� D.V., S�� R.A., S�� M., T��- formation of AA. �� D.H., E�� B.L., E�� D.K., G�� D.C., M�� Th.R., S�� M.A., R�� G.P., Acknowledgement: This work has been carried out with V�� M.D. (2003): J. Agric. Food Chem., 51: support from the European Commission, Priority 5 on Food 4782. Quality and Safety (Contract n°FOOD-CT-2003-506820 [10] M�� M. (2004): J. Biochem. Biophys. Meth., Specific Targeted Project), “Heat-generated food toxicants in press. – identification, characterisation and risk minimisation”. [11] C�� S., A�� M.R., M�� E., O�� M.B.P.P., This publication reflects the author’s views and not neces- F�� M.A. (2003): J. Agric. Food Chem., 51: sarily those of the EC. The information in this document is 6495.

24 Proc. Chemical Reaction in Food V, Prague, 29. 9.–1. 10. 2004