Formation of Acrylamide during Roasting of Coffee Dissertation by MSc. Kristina Bagdonaite was carried out in the period of November 2003 to March 2007 at the Institute for Food Chemistry and Technology, Graz University of Technology supervised by Ao.Univ.-Prof. Dipl.-Ing. Dr.techn. Michael Murkovic.
111
Embed
Formation of Acrylamide during Roasting of Coffee · Formation of Acrylamide during Roasting of Coffee ... 60 Acrylamide formed in asparagine and glucose mixtures ... generated
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Formation of Acrylamide during Roasting of Coffee
Dissertationby
MSc. Kristina Bagdonaite
was carried out in the period of November 2003 to March 2007 at the Institute for
Food Chemistry and Technology, Graz University of Technology supervised by
Ao.Univ.-Prof. Dipl.-Ing. Dr.techn. Michael Murkovic.
6.6 Acrylamide and 3-aminopropionamide formation in a model system . 546.6.1 Preparation mixtures of asparagine with sucrose and glucose . 546.6.2 Sample preparation for optimal heating conditions estimation 556.6.3 Asparagine with ascorbic acid mixture preparation . . . . . . 55
6.7 Heating of pure asparagine . . . . . . . . . . . . . . . . . . . . . . . . 556.7.1 Acrylamide formation from pure asparagine heated at 170 °C
40 Coffee beans roasted under different time and temperature conditions 5441 1-N -(asparaginyl)-5-azido-1,5-dideoxy-D-fructopyranose . . . . . . . . 5742 Reaction of 3-aminopropionamide to sulphonamide . . . . . . . . . . 5843 Derivatization of acrylamide with 2-mercaptobenzoic acid . . . . . . 5944 Typical chromatogram of acrylamide using ion exclusion chromatog-
raphy with UV detection . . . . . . . . . . . . . . . . . . . . . . . . 6045 Typical chromatogram of 3-aminopropionamide analysis . . . . . . . 6146 Typical chromatogram of acrylamide using MS detection . . . . . . . 6247 Typical chromatogram of acrylamide using UV detection . . . . . . . 6448 Typical chromatogram of acrylamide (6 ng/ml) after derivatization
with 2-mercaptobenzoic acid using LC-MS/MS . . . . . . . . . . . . 6549 Typical chromatogram of acrylamide in a roasted coffee sample after
derivatization with 2-mercaptobenzoic acid using LC-MS/MS . . . . 6650 Typical chromatogram of derivatized acrylamide analysis . . . . . . . 6751 Formation of acrylamide in 4 different types of coffee roasted in a
laboratory roaster . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6852 2nd Order regression curve of acrylamide content in coffee beans . . . 7053 Regression surface curve of acrylamide content in coffee beans . . . . 7154 Significant parameters in the formation of acrylamide . . . . . . . . . 7255 Acrylamide content in different coffee beans roasted under standard
62 Acrylamide content (µg/g) in asparagine mixtures with glucose (1:0.5)heated at 250 °C for 1, 3, 4, 5, 7 and 10 minutes . . . . . . . . . . . 80
63 Acrylamide formation from pure asparagine heated at high temperatures 8264 Acrylamide formation in the mixtures of asparagine with ascorbic acid 8465 Acrylamide formation in asparagine and ascorbic acid mixtures heated
at 250 °C and different times . . . . . . . . . . . . . . . . . . . . . . 8466 3-Aminopropionamide formation in the mixtures of asparagine with
a very polar molecule with a molecular weight of 71, a melting point of 84.5 ± 0.3 °Cand a high boiling point of 136 °C at 3.3 kPa [2, 3]. Acrylamide is very soluble in wa-
ter, alcohols, acetone, acetonitrile, slightly soluble in ethyl acetate, dichlormethane,
diethyl ether. It is insoluble in hexane and other alkanes and alkenes. Low, but
significant volatility of acrylamide was observed. It has no significant UV-absorption
above 220 nm and does not fluoresce.
Figure 1: Acrylamide
The amide group is protonated by medium and strong acids. Acrylamide con-
tains a reactive electrophilic double bond and a reactive amide group. It exhibits both
weak acidic and basic properties [4]. Acrylamide can be produced industrially for the
synthesis of polyacrylamide. Polyacrylamide can be used in waste water treatment
as a flocculent, in soil stabilization, in grout for repairing sewers, in the cosmetics,
paper and textile industries [4, 5]. Polymerized acrylamide is also widely used in elec-
trophoresis for protein separation. Acrylamide is described as a neurotoxin, genotoxin
and is probably carcinogenic to humans [5, 6].
Some analytical methods for acrylamide have been reported. The sample prepa-
ration mainly consists of water extraction and a solid phase extraction procedure for
the sample clean-up [7, 8, 9, 10, 11, 12, 13, 14]. Acrylamide can be analyzed by gas
Chocolate milk <25Cocoa powder <100Prune juice 93Chocolate bar <25
plain block <30Nuts Cashew traces
Mixed, salted tracesPeanuts, dry roasted <25
Canned soups <25-50Baked beans canned <25Spanish olives black 345
green <25Meat pie <25Fried rice <25Pizza frozen oven baked <25
microwave 5 min <25Fish crumbed and fried <25
crumbed and oven baked 52Chicken schnitzel <25Hash brown fried 440
oven baked 200-320Beef schnitzel fried and oven baked <25
Surprisingly small amounts of acrylamide are reported in Australian foods (Table
3). The highest concentration of acrylamide is declared only in prune juice, Spanish
black olives, crumbed fish and hash, fried and oven baked as well. Such a huge
difference could be the outcome of not well developed analysis methods or because
only food products that are popular in Australia were chosen and which contain the
lowest amounts of acrylamide. However, potato products, such as French fries or
potato chips are not included.
In the Table 4 data from the United Kingdom are presented. The food infor-
mation sheet [32] says, that the used food samples represent the average UK diet
and sampled foods are prepared according to normal domestic practice. The dietary
exposure estimates showed that cereal-based products and potatoes are the main
9
Table 4: Mean concentrations of acrylamide in UK food [32]Food group Acrylamide, ppbBread 12Miscellaneous cereals 57Meat products 13Fish <5Oils and fats <3Sugars and preserves 23Green vegetables <2Potatoes, made-up 53-112Canned vegetables <5Fruit products <1Beverages <1Nuts <3
source of acrylamide in this European county. Acrylamide was detected in a lot of
products including cereal, meat, sweets, preserves and vegetable products. Though
presenting the quantified results with amounts of acrylamide in everyday’s food, the
Food Standard Agency does not tell people what they should eat but recommends
them to eat all types of foods including cereal, potato products, plenty of fruits and
vegetables avoiding sugar and fatty foods.
It has been noticed, that after a while the acrylamide content in certain food
matrices is getting less. The reduction of this toxin was observed in instant coffee (67
% in one year), roasted coffee (28 % in 7 months), roasted barley (15 % in 9 months)
and cocoa (2 % in 3 months). However, there are also some products (e.g. breakfast
cereals) in which the concentration of acrylamide is not changing over a prolonged
storage period of up to one year [4].
In ground coffee there can be relatively high amounts of acrylamide up to 400
ng/g powder. As acrylamide is a very polar substance, it is not surprising that it is
also detected in large quantities in brewed coffees. In analyzed grounds after brewing
no acrylamide was detected. It seems that all acrylamide available in coffee powder is
transferred to the water where it is quite stable. No significant decrease in acrylamide
levels was observed even after 5 hours of heating [33].
10
4.3 Toxicity of acrylamide
Polymeric acrylamide is nontoxic, but a monomer is neurotoxic to both humans and
laboratory animals. Acrylamide is carcinogenic to laboratory rodents and is described
by the International Agency for Research of Cancer as a probable carcinogen to hu-
mans [5]. Furthermore, acrylamide is not mutagenic in prokaryotic (organisms, that
do not have a real nucleus in their cells) mutagenesis assays, but chronic acrylamide
intake has shown to produce tumors in both rats and mice [34]. Acrylamide’s neuro-
toxicity is characterized by ataxy, distal skeletal muscle weakness and the numbness
of the hands and feet. In the human body acrylamide is oxidized into the epoxide
glycidamide (2,3-epoxypropionamide) via an enzymatic reaction (Figure 2), possibly
involving cytochrome P450 2E1 [4].
Figure 2: Glycidamide formation from acrylamide
Both acrylamide and glycidamide can form hemoglobin adducts [35], but only
glycidamide has been shown to form adducts with DNA amino groups. This fea-
ture of glycidamide implies genotoxicity. Furthermore, high levels of acrylamide can
cause genetic mutations and cellular transformation [34]. Both acrylamide and gly-
cidamide can be detoxified in the cells by glutathione conjugation, or by hydrolysis
[5]. Furthermore, higher acrylamide hemoglobin adducts are detected in smokers,
because acrylamide is also found in tobacco smoke as a result of an incomplete com-
bustion or heating of organic matter [26]. In experiments with rats it was observed
that N -(2-carbamoylethyl)valine adducts were of the magnitude that is similar to the
background level of nonsmoking humans [36]. Furthermore, acrylamide can be ab-
sorbed through the skin, but dermal uptake is only approximately 7 % of oral uptake
[37].
Acrylamide binds to DNA directly via Michael addition reaction in vitro and in
11
vivo [34]. Acrylamide can be excreted from the human body with urine, mostly in me-
tabolized form. N -acetyl-S -(3-amino-3-oxopropyl)cysteine (Figure 3) appears to be a
major metabolite (72 %). The other possible metabolites are glyceramide (hydrolized
glycidamide, 11 %) in Figure 4, glycidamide (2.6 %), N -acetyl-S -(3-amino-2-hydroxy-
3-oxopropyl)cysteine (Figure 5) and N -acetyl-S -carbamoylethylcysteine-S -oxide (14
%) [4].
Figure 3: N -acetyl-S -(3-amino-3-oxopropyl)cysteine
Figure 4: Glyceramide
Figure 5: N -acetyl-S -(3-amino-2-hydroxy-3-oxopropyl)cysteine
Metabolism of glycidamide in humans is mostly via hydrolysis with little via
glutathione conjugation.
12
4.4 Acrylamide formation pathways
There are a few acrylamide formation mechanisms postulated in the scientific re-
ports. Zyzak et al. [38] declare a decarboxylation of the Schiff base pathway,
which is followed by imine’s heterocyclic cleavage or imine’s hydrolysis, when 3-
aminopropionamide is formed and then deaminated to acrylamide. Schieberle and
Granvogl [39] propose the simplest decarboxylation followed by deamination of as-
paragine reaction pathway. Yaylayan et al. [40] show a decarboxylation of the
Schiff base via oxazolidin-5-one intermediate, decarboxylated Amadori product and
β-elimination. Stadler et al. [41] and Yasuhara et al. [42] propose a pathway of
acrylamide formation from fat degradation products when acrylic acid and acrolein
are reacting with amino acid degradation product ammonia. It was reported, that
acrylamide can only be formed from compounds that are naturally present in raw
foods. It cannot be formed from decomposed polymers used for agricultural crops
[43, 44].
Acrylamide is found in large amounts in potato products. It is experimentally
proven that for acrylamide formation in potato products free reducing sugars, mainly
glucose and fructose, have a major influence compared to free amino acids [45, 46, 47].
In experiments the addition of fructose had the strongest effect on the increase of the
acrylamide concentration. Whereas, the addition of asparagine had a rather weak
effect. When asparagine was added in combination with fructose it gave a strongest
response.
Acrylamide is formed in the outer layer of the deep fried potato products (French
fries, crisp, chips). Cooking time and temperature, but not cooking oil type have the
greatest influence on acrylamide formation [48]. With salting or pre-drying, lowering
the pH of the product the acrylamide content during frying can be reduced [4]. For
French fries it is suggested to reduce the frying temperature towards the end of
frying in order to reduce the acrylamide content and still develop a good product
color [49, 50]. Another possibility is to coat the potato slices with proteins, e.g. from
chick peas [51].
13
In contrast to potatoes, the addition of glucose or sucrose to bread dough has a
very small effect [45]. An addition of free asparagine to the flour before the dough
preparation increases the acrylamide contents in baked bread drastically. Therefore
it was concluded, that free asparagine is a limiting factor for acrylamide formation
when we talk about bread and bakery products [52, 53]. A possible way to reduce
acrylamide in bakery products can be the addition of citric acid. It not only reduces
the pH and acrylamide amount in bread, but it also reduces browning, sufficiency of
leavening and affects the taste. The other possibility to reduce acrylamide in bakery
products is to reduce the oven temperature and to increase the baking time [54],
as well as reducing the raising agent ammonium bicarbonate and use ammonium
hydrogencarbonate instead [4].
As lot of experiments show that acrylamide is not formed at temperatures below
120 °C. However it is noticed, that in the heated mixtures of asparagine with sucrose,
a certain amount of water is needed to hydrolize the oligosaccharides to transient
intermediates (glucose and fructose) in order to form the acrylamide [4].
Another product, that is heated to high temperatures to improve its taste and
aroma properties, is coffee. Though coffee beans are roasted at quite high tempera-
tures (220 - 250 �), the acrylamide amounts found in the roasted beans and ground
coffee were reported to be low [6]. There are no significant differences if coffee is
decaffeinated or not. Since consumers are not eating the ground coffee beans, but
prepare a beverage, it is important to calculate acrylamide content not per gram
coffee powder, but per cup. As it was reported [33], during the brewing step of the
coffee beverage, almost all acrylamide present in the coffee powder, is transferred to
the liquid phase of the coffee drink, due to its high solubility in water. Furthermore,
the dietary intake of acrylamide from coffee in northern countries (Norway, Sweden)
can be more than 30 % [55], in Denmark 20 % [56], 36 % in Switzerland [57]. No
data from other countries is available yet.
In coffee, acrylamide is formed in high concentrations during the first minutes of
roasting, resulting in >7 mg/kg. The increase of roasting time leads to the degra-
dation of acrylamide. Kinetic models and spiking experiments with isotope labeled
14
acrylamide showed that >95 % of acrylamide is lost during roasting [3, 54]. However,
the roasting conditions have an important influence on the typical coffee aroma and
taste that are desirable to consumers. Therefore, the optimization of the roasting
conditions with respect to a reduction of the acrylamide formation and maintaining
the product quality has not been realized yet.
Recent reports have announced that acrylamide is not stable in commercial coffee
that is stored in its original container [33].
It is known that the acrylamide formation is favored under low moisture conditions
[4]. It was reported that in a model systems based on fructose the acrylamide content
was increasing with increasing water activity. It is interesting to notice, that in
most experiments fructose with asparagine mixtures were more efficient in acrylamide
formation than mixtures with glucose. Even under anhydrous conditions fructose is
highly reactive and forms higher amounts of acrylamide.
It was recently reported, that the time and temperature of heating have a direct
influence on the acrylamide formation in foods. Acrylamide can already be formed
at 120 °C at a longer time of heating in asparagine mixtures with glucose, whereas
at 160 °C the highest acrylamide concentration was obtained at shorter heating time
[4].
It was reported that the pH has an influence on the acrylamide formation. Adding
some acidifying compound into the potato matrix, at a low pH (<6), a decrease of
acrylamide formation was observed. In contrast, at high pH (∼8) the highest content
of acrylamide formed was detected in potatoes heated to 160-180 °C [58]. It is worth
to notice, that the extraction of acrylamide at pH 2-7.5 gives the same results as a
normal water extraction. But an extraction at a pH >8 gives an increase of acrylamide
recovery and reaches the maximum at about pH 12 [59]. In most scientific works
acrylamide extraction is performed at normal water extraction conditions, and it is
assumed that all the water-soluble acrylamide is bioavailable [59]. By changing the
pH the matrix of the food can be changed and the chemical available acrylamide
dissolves [60]. But as it is known, acrylamide additionally released from a chemically
bound form at the high pH is not bioavailable.
15
4.4.1 The Maillard reaction
The Maillard reaction is named after the first researcher investigating the reaction
of amino acid with carbohydrates, Louis-Camille Maillard (1978-1936). It was then
reported, that amino acids in combination with different carbohydrates react at tem-
peratures above 100 °C forming the typical dark brown color and flavor of cooked
foods [3, 61]. Due to this extremely complex reaction not only desirable color, aroma
and taste properties are gained, but also antioxidant compounds are formed [3].
L.C.Maillard found, that the aldehyde group of an aldose reacts more efficiently
with amino acids than the hydroxyl groups [61]. The melanoidins (reaction products
with a dark brown color) are soluble in the early reaction stage and become insoluble
later. It was found, that the most reactive amino acids are alanine, followed by valine,
glycine, glutamic acid, leucine, sarcosine and tyrosine. Monosaccharides react easily
with amino acids. Sucrose, a non reducing sugar, needs to be hydrolized first to form
the reactive intermediates. Pentoses react faster than hexoses.
The Maillard reaction is devised into three stages. At first, the condensation
reaction of the carbonyl group from a reducing sugar with an amino compound takes
place. In this way the Schiff base is produced. Acid-catalyzed rearrangements give
Amadori rearrangement products, which are not stable above room temperature. In
the reaction of an amino acid with ketose a deoxyaldosylamine is formed by the Heyns
rearrangement [3, 61].
In the second reaction stage, during further enolisation, deamination, dehydration
and fragmentation steps various products including furfurals, furanones, pyranones
and other sugar dehydratation and fragmentation products are formed. In this stage,
amino acids also undergo deamination and decarboxylation through Strecker degra-
dation. α-Hydroxycarbonyls and deoxyosones, intermediates of the Maillard reaction,
and dicarbonyls can act as Strecker reagents and produce Strecker aldehydes. Such
aldehydes can also be formed from a Schiff base. Dehydroascorbic acid, found in
foods, can act as Strecker reagent as well.
In the third stage of the Maillard reaction high molecular mass substances (melanoidins)
16
form during a condensation reaction between carbonyls (especially aldehydes) and
amines [3].
The Maillard reaction was shown to be one of the major pathways in acrylamide
formation both in food matrices and in model systems in the presence of asparagine
and carbohydrates (Figure 6) [38, 45, 62]. Labeling experiments confirm that the
carbon skeleton and the nitrogen of the amino group are derived from asparagine
[38].
Figure 6: Acrylamide formation pathway from asparagine and dicarbonyl (adaptedfrom [63])
For this pathway dicarbonyl compounds (e.g. from glucose) and free amino acids
(e.g. asparagine) form a Schiff base. Further reactions lead to the formation of
3-aminopropionamide and finally acrylamide [38, 63].
It was shown in some experiments that out of twenty amino acids heated with
glucose only asparagine gave significant quantities of acrylamide. In comparison,
however, glucose, fructose, galactose, lactose and sucrose produced similar quantities
of acrylamide. Surprisingly, sucrose, a non-reducing sugar, can produce almost as
17
much acrylamide as some of the reducing sugars in a reaction with asparagine . It is
suggested, that sucrose in real food matrixes can undergo hydrolysis and form glucose
and fructose [3].
During the Maillard reaction a very large number of volatile compounds are
formed. Already more than 550 compounds have been identified. Over 330 com-
pounds found in the reaction systems are volatiles. Most of them give typical flavor
description of different food products [3]. In roasted coffee melanoidins are respon-
sible for the dark color and partly for the aroma. It has been noticed, that coffee
melanoidins can react with volatile compounds and in this way modify the aroma per-
ception in the coffee beverage. Furthermore, the melanoidin spectrum is significantly
influenced by the roasting degree: the darker the beans, the more high-molecular
weight melanoidins can be detected [64].
Maillard reaction some products can be mutagenic, but not carcinogenic [3, 61].
Pyrolyzates of proteins, peptides, and amino acids, especially tryptophan and glu-
tamic acid, coffee beverages prepared in a usual way, black and green tea were re-
ported to contain some mutagenic compounds resulting from the Maillard reaction.
Moreover, mutagenicity appears only above 400 °C and even more strongly at 500-600°C as the pyrolysis temperature [61].
Some compounds showing antioxidant activity were also detected among the Mail-
lard reaction products. These compounds, preventing oxidation of lipids are found
in the melanoidin fraction [61].
4.4.2 Lipid degradation
Another possible pathway for acrylamide formation in heated foods, especially in
deep-fried food products, was suggested to be an oxidative lipid degradation, when
acrolein and acrylic acid - the precursors of acrylamide - are formed [50] and then
they can react with ammonia formed from amino acids (Figure 7) [42], e.g. aspartic
acid [41]. In [4] it is noticed, that in the model system with selected amino acids
and glucose not only aspartic acid can be described as a possible precursor of acrylic
18
acid, but also β-alanine and carnosine (dipeptide). Acrylic acid can be generated
indirectly from other amino acids such as serine and cysteine, forming pyruvic acid,
then reducing it into lactic acid and finally dehydrate into acrylic acid. Free amino
acids in foods such as asparagine, glutamine, cysteine and aspartic acid can be one
of the main sources producing ammonia under thermal treatment of foods [4].
OR
OR
OR
OAcrolein
O
OH
Acrylic acid
NH3
NH2
O
HOOC
NH2
Asparagine
+ Glucoseheat
NH2
OAcrylamide
Figure 7: Acrylamide formation pathways from acrolein and asparagine with sugars(adapted from [62])
Acrylic acid may also appear as an acrolein oxidation product, which can be
formed in the thermal degradation of lipids, either from the oxidation of fatty acids
or from the glycerol moiety [3]. Other sources of acrolein could be amino acids.
A high concentration of acrylamide was found in fried potato products. Experi-
ments showed a slight difference among potato products fried in different oils. It was
observed, that samples fried in olive oil had a higher acrylamide content. Further-
more, the addition of some antioxidants, e.g. rosemary herb, reduced the acrylamide
formation by ∼25 % [62]. In the model system, when asparagine was heated in dif-
ferent oils and fats, it was observed that lard or bovine fat, both containing low levels
of unsaturated lipids, produced lower amounts of acrylamide [4]. The results of these
experiments show, that the higher the degree of unsaturation of the lipids, the larger
the amount of acrylamide is formed. Thus it seems, that acrylamide is not formed
19
Figure 8: Acrylamide formation from asparagine by simple decarboxylation anddeamination reaction [39].
directly from acrolein present in the oil itself. Additionally, by increasing the frying
time the acrylamide content in the samples was also increasing.
The acrylamide content formed in the baked, fried food products depends on
the precursors (amino acids, sugars) amount in the raw material. It was noticed,
that using an industry standard procedure and a frying temperature of 180 °C, the
acrylamide content in potato products can be properly reduced when potatoes with
low sugar content are selected [65].
4.4.3 Decarboxylation and deamination of asparagine
Recently it was reported that acrylamide can be formed from asparagine in absence
of glucose [39]. For this reaction high temperatures (>170 °C) are needed. It was
noticed, that a direct decarboxylation followed by deamination of asparagine can
occur, where reducing sugars do not take place in the reaction (Figure 8).
Furthermore, the amounts of acrylamide were up to 100 times lower in comparison
with model system where glucose was included. This fact helps to understand why
most foods heated at high temperatures contain acrylamide: low amounts of free
asparagine can be detected in almost all kinds of food products [39].
According to Yaylayan et al. [40], the main product is maleimide (Figure 9) and
not acrylamide, when pure asparagine is heated in model system. Maybe because of
the relatively high reaction temperatures (250 and 350 �) asparagine performed fast
intramolecular cyclization and so prevented the formation of acrylamide. Addition-
ally, it was reported that maleimide can be formed also in the presence of reducing
20
sugars.
Figure 9: Maleimide (2,5-pyroldione)
Fumaramic (Figure 10)acid could be one of asparagine degradation products. Its
decarboxylation could also lead to the formation of acrylamide [39].
Figure 10: Fumaramic acid
It seems, that heating of foods at extremely high temperatures (>250 �) for
a shorter time can decrease the amount of acrylamide, because at these tempera-
tures the cyclization of asparagine is preferable and this prevents the formation of
acrylamide [40].
4.4.4 Other precursors for acrylamide formation (3-aminopropionamide)
The direct precursor of acrylamide was shown to be 3-aminopropionamide. The
Strecker alcohol of asparagine (3-hydroxypropanamide) was studied under pyrolitic
conditions [63]. Acrylamide can be easily formed from 3-aminopropionamide by a one-
step dehydration process. As it is mentioned in [63], significant amounts of acrylamide
21
Figure 11: Enzymatic 3-aminopropionamide formation from asparagine (adaptedfrom [66])
in fructose/asparagine mixtures are formed in high quantities at temperatures >180�, whereas significant amounts of acrylamide from 3-aminopropionamide can already
form at 140 °C and the maximum concentration can be achieved at 180 �. It was
also observed, that the thermal degradation of 3-aminopropionamide to acrylamide
under aqueous or low water conditions can already generate at temperatures between
100 and 180 °C [66].
When we look at the asparagine molecule, we can notice, that after we eliminate
NH3 and CO2, we get the acrylamide molecule. In raw materials enzyme decar-
boxylase might generate the biogenic amine (3-aminopropionamide) from asparagine
already at temperatures <100 °C (Figure 11). This reaction does not involve reducing
carbohydrates. The biogenic amine can be formed from the amino acid with pyri-
doxal phosphate as a cofactor. As it is known that enzymatic reactions take place
22
under physiological enzymatic conditions (high water content, warm (10-40 �) tem-
peratures), 3-aminopropionamide formation was especially observed in potatoes [66].
As 3-aminopropionamide is easily deaminated to acrylamide during the heating of
food products, e.g. potatoes, it can be explained why some potatoes during heating
produce more acrylamide though the free amino acid and sugar content in this raw
material is low. Furthermore, the last studies show, that not only potatoes, but also
milk products, e.g. fermented cheese, can contain 3-aminopropionamide and after
heating this can react to form acrylamide [39].
4.5 Michael addition
Losses of acrylamide were reported by a number of researches. This decrease in acry-
lamide amount in processed foods could be a result of evaporation or polymerization
of the monomer. However, it is more likely that acrylamide still reacts with other
components in the food matrix. Such a reaction between acrylamide and nucleophilic
groups, is called Michael addition (Figure 12). Some of such nucleophilic groups could
be amino or thiol, which could be present in free amino acids or as peptides and pro-
teins such as the sulfhydryl group of cysteine, ǫ-amino group of lysine or N-terminal
amino group of proteins.
Figure 12: Michael addition reaction
The Michael addition reaction may be reversible and in certain circumstances it
can lead to the release of acrylamide [3].
23
4.6 Coffee: plants, beans and their production
4.6.1 Differences of Arabica and Robusta
Coffee plants belong to the Rubiacea family, which includes more than 500 genera and
∼8000 species [67]. There are at least 66 species of the genus Coffea L.. Economically
important sorts are Coffea arabica L., Arabica coffee, which represents three quarters
of the world coffee productions, and Coffea canephora Pierre, Robusta coffee, which
makes only one quarter of the world coffee production. Both species are grown
in tropical and subtropical regions: in Central and South America mainly Arabica
is produced, and Africa and South Eastern Asia countries are the main Robusta
production region.
The Arabica and Robusta coffees have different characteristics. The plant (Figure
13), the Arabica coffee tree is 2.5 to 4.5 meters high, is usually cultivated in highlands.
It is resistant to lower humidity conditions and can survive at colder temperatures.
The plant of the Robusta tree is 4.5 to 6.5 meters high, is cultivated more in lowlands.
It tolerates warmer temperatures and higher humidity and is more sensitive to the
cold. At cupping Arabica beans have a good bitter/acid balance and chocolaty to
flowery aroma while the Robusta coffee beans are famous for more bitter taste and
a woody to earthy aroma. That has an influence on the price of the green beans:
Robusta’s price is 20-25 % lower than Arabica’s. In addition, Arabica coffees contain
less caffeine than Robusta (1.2 and 2.2 %, respectively).
The coffee tree in blossom period forms clusters of two to nineteen white flowers.
Fruits are 1.5 cm in diameter and are called ”cherries”, which ripen for seven to nine
months (Arabica coffees) or even nine to eleven months (Robusta coffees). The cherry
has a red or yellow exocarp (skin) when ripe with a gelatinous-pectic mesocarp (the
pulp) of 0.5 to 2 mm thickness, rich in sugars and water, glued over the endocarp
(the parchment), enclosing each seed. The pectins present in the pulp may work as
an energy storage. There are usually two seeds (beans) per cherry, sometimes there
can be only one seed, which is called peaberry. Seeds can vary in size, shape and
24
Figure 13: Coffee plant
density according to growing conditions and genotype.
The cell structure of the beans is characterized by very thick cell walls, which
make raw beans extremely hard.
Amino acids are present in green coffee beans both free (5 % of the total) and
bound to proteins. The free amino acid level depends on maturation. According to
Murkovic et al. [68], the main free amino acid detectable in green coffee beans is
alanine, followed by asparagine. Furthermore, more free amino acids are found in
Robusta green coffee beans.
Carbohydrates are present in green coffee beans both as insoluble and as soluble
polysaccharides, with some arabinose, oligosaccharides, mainly sucrose and traces of
reducing sugars [68, 69]. More sucrose is detected in Arabica (up to 90 mg/g) than in
Robusta beans, but Robusta beans contain more reducing sugars e.g. fructose [68]. In
addition, the glucose content in green coffee beans may influence the cup sweetness.
There are only 0.2-0.3 % of lipids in the green coffee beans present in a waxy layer
surrounding and protecting the beans.
25
4.6.2 Harvesting and processing
Harvesting is done in one of two ways, either by stripping or by picking. In the case
of harvesting by stripping, all cherries (immature, ripe, overripe (raisins), dry) are
stripped together with leaves from the branches at the same time, either by hand or
mechanically. When the picking or finger-picking method is used, only ripe cherries
are hand-picked, and collected in baskets or heavy pieces of cloth laid underneath the
trees. Sometimes techniques that combine stripping and picking can be used, when
only ripe cherries are picked and the other - overripe, green, dry - are collected very
late in the season. Another harvesting method, that is sometimes used, leaves the
cherries on the trees until they are dried and then they are harvested all at the same
time late in the season by shaking the trees.
The crop processing must start as soon as possible after harvesting the cherries
in order to avoid the fermentation and off-flavor formation. The cherries can be
processed either by dry or wet method. Near the equator, only a wet process is
possible because of the rainy climate condition throughout the year. In areas, with
little rainfall and where enough sunshine is ensured, the dry process method must
be used. Subtropical areas, where the rainy and dry season are well defined, both
methods (dry or wet) can be used.
Beans processed with the dry method are sun-dried on patios. After the cherries
are brought in from the fields, they are first cleaned from impurities by compressed
air and by washing. By using water the leaves, wood sticks and soil particles can be
separated from the beans. Also cherries can be separated into two categories: heavier
ripe cherries and lighter overripe ones. After twelve to twenty days of drying the
beans finally contain 12 % humidity and are ready for roasting. Coffees processed by
this method are so-called natural coffees, giving the cupping a full body and a mild
aroma.
The wet method can only be used with finger picked cherries and when there
are only a few or no not yet ripe cherries in the harvest. The coffee beans are
separated by removing their skin mechanically and the mucilage is washed away after
26
the fermentation step.
In the beginning the cherries are washed in large water tanks, where heavy (ripe)
and light (immature) cherries are separated. After the cherries move through the
depulping machine, which removes the skin and some adhering pulp, they remain
covered with a layer of 0.5-2 mm thick mucilage. Due to naturally present microor-
ganisms, the beans ferment now for 18-36 hours, depending on ambient temperature,
wether they are stored in brick or concrete tanks, or wether they are immersed in
water (so-called wet fermentation) or not (so-called dry fermentation). Fermentation
ends when the parchment loses the slimy layer of the mucilage. After the beans are
washed and dried to a humidity level of ∼12 % they are ready for roasting. When
beans are processed by the wet method, they are called washed coffees, resulting in
a highly aromatic cup with a fine body and an acid live aroma.
The third method, that can be used, is a semi-dry process [70]. It is a mixture
between dry and wet processing methods where the fermentation step is omitted.
The beans, when this method is used, are washed and depulped. The remaining
mucilage is not fermented, and the beans are dried in the sun. After the beans are
dried to ∼12 % of humidity, the mucilage is mechanically removed and the beans are
transported to a roasting facility.
Usually the beans are electronically sorted before the roasting to avoid that defect,
fungal contaminated beans or other beans with unwanted properties which could have
an undesirable effect on the cup quality, are roasted as well. It was observed, that
so-called black and sour beans roast to a smaller degree, thus reducing the beverage
quality [71]. Interestingly to notice that natural coffees regularly have more defects
compared to the washed ones, but the natural coffee beans are easier to sort out [67].
4.6.3 Coffee roasting
The roasting process takes place at quite high temperatures: It can range from 240°C up to 300 °C for industrial roasting [72]. For laboratory bean roasting 24 minutes
at 220-230 °C could be estimated as optimal conditions for the acceptable sensory
27
properties for the coffee beverage [73].
Roasting induces several visible changes in color, texture, density and size [74, 75,
76]. Also the beans lose up to 8 % in dry weight, mainly because of the loss in sugars
taking place in the Maillard reaction and pyrolysis [77]. It was observed, that coffee
beverages show some antioxidant properties [78, 79]. Especially green coffee beans
are rich in antioxidative compounds (e.g. chlorogenic acid), which during roasting
are almost lost [80]. The roasting process for the coffee beans has some advantages:
not only the pleasant taste and aroma are formed, but also the natural contaminant
ochratoxin A can be eliminated [81, 82].
The roasting process can be divided into three phases. The initial is a drying phase
when the moisture is eliminated from the coffee beans, keeping the bean temperature
at around 100 �. This phase lasts only a few seconds since the green coffee beans
contain no more than 12 % of humidity.
The second phase is the actual roasting step. In this phase at ∼170 °C complex
exothermic pyrolytic reactions take place. Large quantities of CO2, water and volatile
substances are released. At temperatures >120 °C the Maillard or non-enzymatic
browning reactions take place, which are followed by Streckers degradation at ∼160°C. Sugar and minor lipid degradation reactions also take place at the temperatures
close to 180-200 �. And of course at such high roasting temperatures intermediate
decomposition products interact as well.
The third phase is the cooling phase. The beans are rapidly cooled using cold air
or water as a cooling agent. Coffee beans can be roasted to a different roasting degree
which depends on the consumers taste. It is known that light roasting is prefered
in northern countries whereas dark or very dark roasting is preferred in southern
regions.
28
5 Purpose of the study
Because acrylamide is very small polar molecule, and coffee in comparison is a com-
plex enough matrix, methods of extraction and clean-up, followed by analytical ones
were established.
In order to understand the differences between the coffee types better, the follow-
ing successive steps had to be taken:� Coffee was roasted to different roasting degrees, using a small scale laboratory
roaster.� The influence on acrylamide formation in coffee roasted under standard roasting
conditions was studied.� Coffee was roasted at different time and temperature conditions in order to
observe the acrylamide formation kinetics.� Extraction and a liquid chromatography with fluorescence detection analytical
method was established in order to determinate 3-aminopropionamide in green
and roasted coffee beans
With a model system imitating similar to acrylamide formation reaction condi-
tions the following steps were taken:� The influence of time and temperature, molar differences on acrylamide and 3-
aminopropionamide formation kinetics in the asparagine mixtures with sucrose,
glucose was observed.� Optimal formation conditions for acrylamide and 3-aminopropionamide in the
anhydrous mixtures of asparagine with glucose were defined.� Pure asparagine was heated in order to determinate acrylamide or other possible
compounds that have formed.
29
� Asparagine mixtures with ascorbic acid were heated in order to observe other
acrylamide formation theories.� Amadori compound was heated in order to review an acrylamide formation
pathway during the Maillard reaction.
30
6 Materials and Methods
6.1 Chemicals and solvent
Water was distilled twice and further purified using a water purification system
First, green coffee beans were ground with an analytical mill and second with
a ball mill in order to get a fine powder. 100 to 500 mg of the fine powder were
balanced into 14 ml centrifuge tube, 7 ml of 0.1 M HCl were added and mixed. After
10 minutes of ultrasonic treatment and centrifugation at 4000 rpm for 30 minutes 5
ml of aliquot were transferred to a 10 ml flask. The pH was adjusted to ∼7 with
0.25 M NaHCO3. Then the flask was filled with water to the mark. Supernatant
was filtered and derivatized with dansyl chloride. Analysis for 3-aminopropionamide
was performed by HPLC with fluorescence detection. The standard solutions of
3-aminopropionamide in 0.25 M NaHCO3 (200, 400 and 600 ng/ml) were used as
external standards.
6.5.2 3-aminopropionamide in heated coffee
First of all 10.0 g of Liberia Robusta (Coffea canephora, dry-processed) green beans
were roasted in an oven at 200, 220 and 240 °C for 5, 10 or 15 minutes (Figure 40).
Secondly Liberia Robusta (Coffea canephora, dry-processed) and Indian Plantation A
Arabica (Coffea arabica, washed) were heated at 150 and 170 °C for 7 minutes. Before
roasting, the glass dishes were preheated in an oven for 10 minutes. After roasting the
samples were immediately put on ice for 15 minutes. After cooling the heated coffee
beans, they were ground in an analytical mill and prepared for HPLC-FL analysis
according to the same procedure used for green coffee.
53
Figure 40: Coffee beans roasted under different time and temperature conditions
6.6 Acrylamide and 3-aminopropionamide formation in a model
system
6.6.1 Preparation mixtures of asparagine with sucrose and glucose
Asparagine and sugars in molar ratio of 1:0.5, 1:1 or 1:1.5 were dissolved in 20 ml of
purified water in a round flask. Samples were frozen (-20 °C) overnight and freeze-
dried to get a fine dry powder. Approximately 50 mg of the freeze-dried asparagine,
sucrose or glucose mixtures were heated in 4 ml vials at 130, 150, 170 and 190 °C for
1 to 30 minutes. After heating the samples were cooled for 40 seconds in the air (20°C) and 15 more minutes on ice. The heated samples were dissolved in 3 ml of 0.25
M NaHCO3 (pH ∼8). After the aliquot had been sonicated for 10 minutes 500 µl of
the solution were centrifuged for 5 minutes, diluted if necessary 10 times and then
derivatized with dansyl chloride or analysed for acrylamide by LC-UV.
54
6.6.2 Sample preparation for optimal heating conditions estimation
Approximately 50 mg of the freeze-dried asparagine with glucose (molar ratio 1:0.5
and 1:1) powder were taken into 4 ml HPLC heated to a provided temperature for
a certain time. After heating the vials were cooled for 40 seconds in the air and
another 15 minutes on ice. After that, the heated each mixture was dissolved in 3 ml
of purified water, it was sonicated for 10 minutes, centrifuged for 5 minutes. Before
the HPLC-UV analysis the aliquot was diluted with purified water 10 times.
6.6.3 Asparagine with ascorbic acid mixture preparation
Asparagine and ascorbic acid in molar ratio of 1:0.5, 1:1 or 1:1.5 were dissolved in
20 ml of purified water in a round flask. Samples were frozen (-20 �) overnight and
freeze-dried to get a homogenious powder. Approximately 50 mg of the freeze-dried
asparagine and ascorbic acid mixtures were heated in 4 ml vials at 150, 170, 190,
210, 230 and 250 °C for 1 to 15 minutes. After heating the samples were cooled
for 40 seconds in the air (20 �) and 15 more minutes on ice. The heated samples
were dissolved in 3 ml of purified water. After the aliquots had been sonicated for
10 minutes 100 µl of the solution were added to 900 µl of 0.25 M NaHCO3 and then
taken for 3-aminopropionamide analysis. The rest of aliquot was derivatized with
2-mercaptobenzoic acid and analysed for acrylamide by using LC-MS.
6.7 Heating of pure asparagine
6.7.1 Acrylamide formation from pure asparagine heated at 170 °C for0-24 minutes
Pure asparagine was heated in 2 ml HPLC vials in a heating block. Vials were
put at once all together (screw caps not totally closed). After heating for a certain
amount of time (0, 3, 6, 9, 12, 15, 18, 21, 24 minutes), each vial was taken out
55
of the heating block, tightly closed and kept in the air for 40 seconds, and put on
ice for another 15 minutes. The heated asparagine samples were dissolved in 1.5
ml purified water, transferred to the 2 ml centrifuge tubes and centrifuged for 5
minutes. The samples were prepared for HPLC-FLD (3-aminopropionamide) and
LC-MS (acrylamide) analysis.
For HPLC-MS analysis 50 µl of the centrifuged sample were purified by SPE SAM
OASIS (Waters, Milford, MA, USA). The cartridges were preconditioned with 2 mL
MeOH and 2 ml H2O. After SPE purification 13C3-labeled acrylamide (final conc.
100 ng/ml) was added.
6.7.2 Maleimide formation at 170 °CPure asparagine was heated for 20 minutes at 170 °C in a 2 ml HPLC vial. Then it
was dissolved in 1.5 ml of purified water. The aliquot was centrifuged and analysed by
HPLC-MS. Analysis was carried out with SIM mode, we were looking for acetamide
tector was used. For analysis Synergi 4 µ Polar-RP 80A, 150 × 4.60 mm i.d. (Phe-
nomenex) analytical column with precolumn 1 × 4.6 mm (Phenomenex) was used.
As a mobile phase water was used. The column temperature was maintained at 20°C, flow rate 1 ml/min, post time was set to 5 minutes, injection volume 10 µl. De-
tection was performed at a wavelength of 210 nm. The acrylamide concentration of
the samples was calculated by external calibration. For this purpose the acrylamide
aqueous standard solutions in the range of 10-10 000 ng/ml were prepared. The limit
of detection (LOD) was determined as 58 ng/ml, the limit of quantification (LOQ)
as 105 ng/ml, RSD of 0.7%.
The typical chromatogram of the acrylamide analysis is given in Figure 47.
6.10.6 HPLC-MS operating conditions for acrylamide derivative analysis
The analysis of acrylamide derivatives was developed according to the method re-
cently described in the literature [16]. Since we could not apply this method di-
rectly in our laboratory, the method was modified and improved with help of Dra.
M.T.Galceran and her research team at University of Barcelona, Department of An-
alytical Chemistry.
For this purpose derivatized acrylamide standard solutions in the range of 0.6-622
63
Figure 47: Typical chromatogram of acrylamide using UV detection
ng/ml and 62.5 ng/ml of d3-acrylamide were prepared. The coffee beans were roasted
at 240 °C for 5, 10 and 15 minutes, prepared according to the method described in 6.4
and derivatized with 2-mercaptobenzoic acid according to the method described in
6.9.2. Before the analysis by LC-MS/MS the solutions were diluted 1:50 with purified
water.
For LC-MS/MS an Agilent HP 1100 equipped with vacuum degasser, quarternary
pump, autosampler coupled to a PE Sciex API 3000 (Applied Biosystems, Foster City,
CA, USA) equipped with an electrospray (ES) as an ionization source and a triple
quadrupole as an analyzer was used. The optimal ionization source working param-
eters were: nebulizer gas 10 a.u.; curtain gas 12 a.u.; vaporizer temperature 400 °C;
electro spray voltage (ion spray voltage) 5.5 kV; declustering potential 50 V. The
data acquisition was performed using selected reaction monitoring (SRM), using as
precursor ion the protonated molecule [M + H]+ and monitoring two product ions,
the most abundant product ion for quantitative purposes and another one for confir-
mation. The transitions precursor→product ion used for acrylamide derivative were
64
m/z 226→191 (quantitative analysis), 191→163 (confirmatory analysis) and while
for deutereted acrylamide derivative were m/z 229→211, 229→193 and 229→194.
The chromatographic separation of acrylamide was carried out by reverse-phase
liquid chromatography using a Waters Symetry C8 150 × 2.1 mm i.d. 5 µm particle
size analytical column, the eluent composition was 30 % acetonitrile and 70 % acetic
acid (0.1 %) aqueous solution. Flow rate 0.3 ml/min was performed in isocratic
elution. The sample injection volume was 10 µl.
Calibration standard solutions of derivatized acrylamide were prepared and anal-
ysed by LC-MS/MS (Figure 48). The chromatogram of the derivatized acrylamide
in roasted coffee is given in Figure 49.
An important signal suppression was observed due to the matrix effect in the
ionization process. To improve the signal it was necessary to dilute the sample with
purified water 1:50.
Figure 48: Typical chromatogram of acrylamide (6 ng/ml) after derivatization with2-mercaptobenzoic acid using LC-MS/MS
For LC-MS analysis in our laboratory an Agilent HP 1100 equipped with vacuum
degasser, quarternary pump, autosampler and MSD was used. The chromatographic
65
Figure 49: Typical chromatogram of acrylamide in a roasted coffee sample afterderivatization with 2-mercaptobenzoic acid using LC-MS/MS
separation of acrylamide was carried out using a Phenyl-Hexyl 150 × 3 mm i.d. 3
µm particle size analytical column, the eluent composition was 30 % acetonitrile and
70 % acetic acid (0.1 %) aqueous solution. The analytical column temperature was
25 °C. Flow rate 0.4 ml/min was performed in isocratic elution. The sample injection
volume was 3 µl. Detection was performed at a MS positive (API-ES) mode, SIM at
m/z 226 (for acrylamide) and 248 (for acrylamide and Na+ adducts).
The acrylamide concentration of the samples was calculated by external calibra-
tion. For this purpose the acrylamide aqueous standard solutions in the range of
10-5000 ng/ml were prepared and derivatized. The limit of detection (LOD) was
determined as 157 ng/ml, the limit of quantification (LOQ) as 270 ng/ml, RSD of
5.6% using software Validata (version 1.01).
The typical chromatogram of the derivatized acrylamide analysis is given in Figure
50 .
66
Figure 50: Typical chromatogram of derivatized acrylamide analysis
67
7 Results and Discussion
7.1 Coffee roasting in a laboratory roaster
Four different types of coffee were used for this experiment: Cameroon Robusta (low
quality), Santos Brazil NY 2 17/18 TOP Italian preparation, Arabica (low quality),
Colombian Excelso Centrals mild, Arabica and Uganda Organico Biocoffee, Arabica
(high quality). Cameroon Robusta and Santos Brazil Arabica were roasted according
to programmes 4, 6, 8 and 10. Both Colombian Excelso and Uganda Organico Bio-
coffee were roasted using programme 6. 80 g of each coffee green beans were taken to
the laboratory roaster. 5.0 g of each roasted coffee were taken for acrylamide analysis
performed by liquid ion exclusion chromatography, and the rest of coffee was put in
amber glass bottles, flushed with nitrogen and stored in the freezer (-20 �). Green
coffee beans (50.0 g) were prepared for storage in the same way.
Figure 51: Formation of acrylamide in 4 different types of coffee roasted in a labora-tory roaster
Roasting programme 6 is usually used for common drinking coffee. During roast-
68
ing the coffee the temperature in a laboratory roaster is high, but normally not above
250 �. Our experiment showed significant difference of acrylamide concentration in
different coffee beans roasted in a laboratory roaster. Cameroon Robusta showed a
significantly larger amount of acrylamide formed during roasting (Figure 51), while
in other types of coffee beans - Arabica, both washed and not washed - concentration
of acrylamide detected was 7.6 (Santos Brazil) to 10.5 (Colombian Excelso) times
less. Furthermore, Cameroon Robusta coffee beans roasted according to programme
4 and 8 had a higher acrylamide concentration compared to Santos Brazil.
Commercially available coffee (Lavazza 100 % Arabica, Torino, Italy) was analysed
as well. Additional standards of acrylamide solutions (50 and 100 ng/ml) were used
for this analysis, and the concentration of this food toxin was calculated from the
calibration curve. We have detected 191 ng/g acrylamide in analysed coffee powder.
7.2 Coffee heated in a thermostatic oven
Cameroon Robusta coffee beans were chosen for this experiment. 10.0 g of beans
were roasted in an oven at 180, 200, 220 °C for 1, 3 and 5 minutes and at 220, 240,
260 °C for 5, 10 or 15 minutes. Before roasting, the glass dishes were preheated in
an oven for 10 minutes. After roasting the samples were immediately put on ice for
15 minutes. The cooled coffee was ground in a coffee grinder and the samples were
prepared for ion exclusion with UV detection chromatographic analysis.
Table 8: Concentration of acrylamide in Cameroon Robusta, heated in an thermo-static oven, ng/g.
Temperature � Time, min5 10 15
220 463 250 0488 158 0
240 191107110
260 239 0 0
69
Figure 52: 2nd Order regression curve of acrylamide content in coffee beans
In our experiments we detected the highest concentration of acrylamide in beans
roasted at 240 °C for 5 minutes (Table 8). Also in coffee beans, roasted at 220 and
260 °C for 5 minutes we detected higher amounts of acrylamide than in those roasted
at 10 or 15 minutes (Figure 52). With the increase of roasting time we noticed a
acrylamide decrease at all temperatures. Figure 52 shows how acrylamide is formed in
coffee beans heated at temperatures between 220 and 260 �. The largest acrylamide
amount forms during the first five minutes and lower temperature (220 �). Heating
at higher temperatures and for longer (5, 10 minutes) time acrylamide concentration
in the coffee beans is decreasing. The acrylamide formation and elimination processes
are faster at higher temperatures than at lower ones.
In the experiment, when coffee beans were roasted at 180, 200 and 220 °C for
1, 3 and 5 minutes (Table 9) the highest acrylamide amount was detected in beans,
heated to 220 °C for 5 minutes. With the increase of the roasting time at all temper-
atures (180, 200, 220 �) we noticed an increase of acrylamide formation. At higher
temperatures (220 �) this increase was more rapid. We observed the acrylamide
increase in the samples, roasted for longer time, e.g. 5 minutes. With the increase
of temperature we noticed a rapid acrylamide increase in the samples. It seems, that
the highest acrylamide amount forms in the first 5 minutes of coffee roasting and
later it decreases. The higher the roasting temperature and the longer roasting time,
70
Figure 53: Regression surface curve of acrylamide content in coffee beans
the faster the acrylamide degradation is in coffee samples.
In our roasting experiments we analysed time and temperature influence on the
acrylamide formation. According to our results (Figure 53) both time and tempera-
ture conditions are significant for the formation of the toxicant.
In the presented coffee roasting experiments we could analyse time and tempera-
ture parameter influence on the acrylamide formation. According to our results we
can affirm that both time and temperature conditions have influence on the forma-
tion of the toxicant. It seems that acrylamide is formed in coffee during the first 5
heating minutes. After that we noticed an acrylamide decrease in our samples. The
higher the temperature and the longer the time of the roasting process, the faster
acrylamide degradation can be observed in the coffee.
Table 9: Concentration of acrylamide in Cameroon Robusta, heated in an thermo-static oven, ng/g.
Figure 53 shows how acrylamide is formed in coffee beans heated at temperatures
71
Figure 54: Significant parameters in the formation of acrylamide
between 220 and 260 �. The largest acrylamide amount forms in first five minutes
and lower temperature (220 �). Heating at higher temperatures and for longer (5,
10 minutes) time acrylamide concentration in the coffee beans is decreasing. The
acrylamide formation and elimination processes are faster at higher temperatures
than at lower ones.
Figure 54 shows the significance of parameters in formation of acrylamide. In
our experiments both time and temperature were significant for the formation of
acrylamide.
7.3 Standard condition roasting
In this experiment experiment 19 types of green coffee beans from different regions
of the world (from Africa (n = 5), Asia (n = 6), Oceania (n = 2), South America
(n = 1), Central America (n = 5) were roasted and analysed for acrylamide. Green
coffee beans are described in Table 6. The green coffee beans were heated in an oven
under standard conditions. Coffee samples of 10 g each were heated at 240 °C for 7
minutes in glass dishes. Glass dishes were preheated for 10 minutes; after heating the
dishes with coffee samples were cooled for 15 minutes on ice. Then the beans were
ground and prepared for analysis by LC-MS.
72
Figure 55: Acrylamide content in different coffee beans roasted under standard con-ditions
When we analysed 19 coffees from different parts of the world, roasted under stan-
dard roasting conditions (240 °C for 7 minutes), we noticed a significant difference
between Robusta and Arabica coffees (Figure 55). Exact values of this experiment are
shown in Table 10. Most of our studied Arabicas were washed and rather high qual-
ity with exception for coffees from Tanzania and one coffee from Kenya. Our results
showed no big difference in acrylamide amount formed in Arabica coffees, whereas
acrylamide amounts in Robusta coffees were higher than in all tested Arabicas. Ac-
cording to our results it seems, that the growth area of coffee beans did not have a
significant influence in acrylamide formation. In coffee beans from Asia we noticed
lower acrylamide amounts in dry-processed and semi-washed beans, whereas washed
Arabica beans and especially monsooned ones had higher acrylamide amounts. In-
donesian coffees had similar acrylamide content, though they were processed by dif-
ferent methods. Also in Robusta dry-processed coffee beans we noticed a lower AA
amount than in washed ones.
73
Table 10: Coffee roasted under standard conditionsCoffee Acrylamide formed, ng/gRobusta Indian Parchment 762Robusta Vietnam 653Indian Monsooned Aspinwalls Malabar AA 575Indian Plantation A 401Indonesian Sumatra Lintong 301Indonesian Sulawesi Kalossi 306Tansania Arabica 354Ethiopian Sidamo Yirgamo Grade 2 425Zambia AA 331Kenia washed 531Kenia A/A 299Nicaragua Talia Extra 433Guatemala SHB 374Mexico Maragogype 363Mexico Altura 380Costa Rica Tarazzu 310Honduras 307Papua New Guinea Sigri C 352Java WIB1 Jampit Gr1 357
It seems, that neither place of origin nor processing method has significant influ-
ence on the acrylamide formation. We can conclude, that medium roasted coffee has
the highest amount of acrylamide and it can be reduced by roasting the coffee beans
to a darker color, allowing the coffee to roast for longer time.
7.4 Acrylamide and 3-aminopropionamide formation in a modelsystem
As it was recently reported 3-aminopropionamide seems to be an important precursor
for acrylamide formation. In our study we prepared asparagine and sugars (sucrose,
glucose) mixtures of different molar ratios, heated and analysed them for acrylamide
and 3-aminopropionamide.
Asparagine mixtures with sugars were heated at 130, 150 and 170 °C for 7 minutes.
In this experiment we could detect both 3-aminopropionamide and acrylamide in high
amounts (Figure 57 and Figure 56).
74
Figure 56: Acrylamide in heated asparagine and sucrose and asparagine and glucose(molar ratio 1:0.5, 1:1 and 1:1.5) anhydrous mixtures at 130, 150 and 170 �
Acrylamide amount in heated asparagine mixtures with sucrose or glucose was
detected much higher than 3-aminopropionamide amount (Figure 56). In asparagine
mixture with sucrose acrylamide was detected already at 150 �. With an increase
of temperature we noticed an increase of acrylamide. Furthermore, the highest con-
centration of acrylamide was detected in heated asparagine samples with sucrose or
glucose at 170 �.
Asparagine mixture with glucose had a higher capacity already at lower tem-
peratures (130 �) to form 3-aminopropionamide and acrylamide. We detected the
highest amount of 3-aminopropionamide in the mixtures heated at 150 �. It seems,
that the molar ratio of asparagine to glucose did not have a significant influence on
the 3-aminopropionamide formation. Whereas in the asparagine mixtures with su-
crose at high temperatures (170 �) we noticed a decrease in 3-aminopropionamide
formation when the sucrose amount in the mixture was increasing. At 150 °C we no-
ticed the opposite effect. We did neither detect 3-aminopropionamide nor acrylamide
in the asparagine and sucrose mixtures heated at 130 °C. This temperature is too low
75
Figure 57: 3-Aminopropionamide in heated asparagine and sucrose and asparagineand glucose (molar ratio 1:0.5, 1:1 and 1:1.5) anhydrous mixtures at 130, 150 and170 �.
for sucrose to form necessary reaction products. In this experiment we also noticed
an acrylamide increase with increasing the heating temperature. Furthermore, the
amount of acrylamide formed in the mixture was similar to 3-aminopropionamide
amount, especially in the mixtures heated at 170 �.
3-aminopropionamide was detected in heated anhydrous mixtures of asparagine
with either glucose or sucrose. The model reaction was carried out in a temperature
range from 130 to 170 �. In the model reaction with glucose 3-aminopropionamide
was formed starting at 130 �, whereas with sucrose 3-aminopropionamide was only
found in the reaction carried out at 170 °C (Figure 57).
As we know, for the formation of 3-aminopropionamide reducing sugars such as
glucose or fructose are needed. As sucrose fragments into glucose and fructose at the
temperature above 170 �, 3-aminopropionamide was not detected at temperatures
below 170 �. Furthermore, in heated mixtures with an increasing molar ratio of
76
asparagine to glucose at 130, 150 and 170 °C and sucrose at 170 °C the formation of
3-aminopropionamide decreased. In addition, in the asparagine samples with glucose,
heated to 150 °C the highest concentration of 3-aminopropionamide was detected.
7.4.1 Optimum time and temperature conditions for 3-aminopropionamideformation
We noticed, that the highest amount of 3-aminopropionamide was formed in as-
paragine and sucrose mixture 1:0.5. The mixture was heated to 130, 150, 170 and 190°C at 7 minutes. The data showed, that the highest amount of 3-aminopropionamide
was detected in the mixture, heated to 170 °C (Figure 58). At this temperature we
heated the asparagine and sucrose mixture for 1, 3, 7, 10, 15 and 20 minutes. At 7
minutes of heating we have detected the maximum amount of 3-aminopropionamide.
It seems, that 3-aminopropionamide is rapidly forming to the highest possible amount
in the first 5-7 minutes and after that it is eliminated by probable transformation into
acrylamide.
7.4.2 Optimal heating time and temperature conditions for acrylamide
We have performed a similar experiment as described in 7.4.1 to estimate the opti-
mum time and temperature conditions for acrylamide formation. For this purpose as-
paragine and glucose mixtures (molar ratio 1:0.5 and 1:1) were prepared. Asparagine
and sugar were dissolved in 60 ml of purified water and freeze-dried. Approximately
50 mg of the fine powder were taken for one sample. The asparagine and glucose
freeze-dried mixtures were heated at 150, 170, 180, 190 and 200 °C for 5 and 7 min-
utes (Figure 59). According to the results shown, the mixture of asparagine and
glucose (1:0.5) was chosen for the next experiments in order to determine optimal
time and temperature heating conditions. As it can be seen in asparagine and glucose
mixture (1:0.5) the acrylamide amount formed during the heating was much higher.
In the next step, asparagine and glucose (1:0.5) freeze-dried mixture was heated
77
Figure 58: 3-Aminopropionamide in heated asparagine and sucrose 1:0.5 anhydrousmixtures at 130, 150, 170 and 190 �
Figure 59: Acrylamide content in asparagine with glucose mixtures (molar ratio 1:0.5and 1:1) heated to different temperatures for 5 and 7 minutes
78
Figure 60: Acrylamide formed in asparagine and glucose mixtures (1:0.5) heated atdifferent temperatures for 5 minutes
for 5 minutes at different temperatures: 190, 200, 210, 220, 230, 240, 250 and 260°C (Figure 60). As it is shown in the Figure 60, we cannot distinguish the optimal
heating temperature because of a lack of obvious extreme value. But because of the
lack of sample for analysis, two heating temperatures (210 and 250 �) were chosen
for comparison.
Figure 61 shows the acrylamide amount in asparagine and glucose mixture (1:0.5)
heated to 210 °C for 1, 3, 5, 7, 10, 15, 20 and 30 minutes. The highest concentration
of acrylamide was observed in the sample, heated for 7 minutes. After 15 minutes
the acrylamide amount in the samples was below the limit of detection.
In Figure 62 values of acrylamide formation in the asparagine and glucose mixture
heated at 250 °C are shown. It seems that acrylamide at such a high temperature
forms extremely quickly. In this experiment the highest acrylamide concentration
was observed already after 1 minute of heating.
In our experiments with a model system (asparagine mixtures with sucrose and
glucose) we observed acrylamide and its potential precursors 3-aminopropionamide
79
Figure 61: Acrylamide formation in asparagine mixtures with glucose at 210 °C(molar ratio 1:0.5)
Figure 62: Acrylamide content (µg/g) in asparagine mixtures with glucose (1:0.5)heated at 250 °C for 1, 3, 4, 5, 7 and 10 minutes
80
formation. It seems, that 3-aminopropionamide needs lower temperatures to obtain
a maximum amount, whereas acrylamide can reach a maximum concentration in the
mixture heated to rather high temperature (250 �). Both substances are rapidly
formed in the first minutes of heating and with increasing of the heating time the
amount of both 3-aminopropionamide and acrylamide is decreasing. However, acry-
lamide formation is connected to a heating temperature value: the higher the heating
temperature, the shorter time is needed to achieve a maximum acrylamide concen-
tration in the mixture. Furthermore, in the asparagine and glucose mixtures both
3-aminopropionamide and acrylamide can be formed already in the mixtures heated
to 130 �.
7.5 Heating of pure asparagine
Heating of asparagine to 170 °C for up to 24 minutes did not result in a formation of
neither acrylamide nor 3-aminopropionamide. In the sample, heated at 170 °C for 20
minutes maleimide (2,5-pyroldione) was detected. (Figure 9). The same results were
obtained by Yaylayan et al. [40], who suggested that carbohydrates or carbohydrate
degradation products are necessary for acrylamide formation from asparagine. The
acrylamide formation pathway from decarboxylation of Schiff base which leads to the
decarboxylation of Amadori products is preferred by most scientists, because as the
experimental data show when asparagine is pyrolyzed in the absence of carbohydrates,
maleimide formes avoiding therefore acrylamide formation.
Surprisingly, we could detect acrylamide in pure asparagine samples, heated at
high temperatures (210, 230 and 250 °C) for 2, 5 and 10 minutes (Figure 63). We have
detected this toxin in the samples, heated at 230 °C for 5 and 10 minutes (with the
increase of heating time, acrylamide concentration was increasing) and in the samples
heated at 250 °C for 2 and 5 minutes. At 250 °C heated samples had more acrylamide
formed than the ones heated at 230 °C. However, the concentrations detected in this
experiment were extremely small (only 2-6 µg/g) in comparison to other previous
experiments, when asparagine was reacting with carbohydrates or ascorbic acid.
81
Figure 63: Acrylamide formation from pure asparagine heated at high temperatures
This experiment can prove, that acrylamide can be formed when pure asparagine
is heated at high temperatures by simple decarboxylation and deamination reaction
(Figure 8). However, the amount of acrylamide formed is very small and this reaction
can not be accepted as a main pathway of acrylamide formation.
7.6 3-Aminopropionamide in coffee
No 3-aminopropionamide was found in green coffee beans. 3-aminopropionamide
can form in raw food stuffs, when enzymatic reaction takes place. According to
Schieberle [66], 3-aminopropionamide can form in a biochemical pathway, when the
decarboxylation of asparagine to 3-aminopropionamide takes place with pyridoxal
phosphate as co-factor. In this reaction the enzyme decarboxylase is needed (Figure
11). In our study we could not detect 3-aminopropionamide in green coffee. Maybe
82
because of disadvantageous conditions of the matrix, where humidity is no more than
6 % and enzymatic reactions do not take place.
We could not detect 3-aminopropionamide in coffee beans roasted at 150, 170,
200, 220, 240 °C for different times neither.
7.7 Heated mixtures of asparagine with ascorbic acid
It was recently delivered that ascorbic acid also known as vitamin C can significantly
reduce acrylamide formation in French fries [87]. Since it is known that vitamin
C can be found in all fresh vegetables including potatoes, it could be considered
as a natural inhibitor for acrylamide formation. In our experiments we tried to
investigate if asparagine and ascorbic acid alone under non-aqueous conditions can
be a substratum for acrylamide formation.
We chose different molar ratios of asparagine to ascorbic acid and heated the
mixtures at temperatures up to 150-250 °C (Figure 64).
Our experiment showed, that at lower temperatures (150, 170 �) there was no
acrylamide detected in the mixtures. Only at 190 °C when the asparagine and vita-
min C molar ratio was 1:1.5 after 5 minutes of heating we detected up to 20 µg/g
acrylamide. With increasing the heating temperature acrylamide concentration was
increasing (heating time 5 minutes) and at 250 °C it reached 50 µg/g. However, after
10 minutes of heating acrylamide concentration was decreasing. In comparison to
higher ascorbic acid amount in the mixture, when asparagine molar ratio to ascorbic
acid was 1:0.5 we observed much lower amounts of acrylamide formed: not more than
26 µg/g. Interesting results brought us the mixture of asparagine and ascorbic acid,
when molar ratio was 1:1. We could detect acrylamide only in the mixture heated
at 230 °C for 5 minutes. In general it is obvious, that acrylamide in asparagine and
ascorbic acid mixtures is formed at relatively high temperatures, short 5 minutes-
time (Figure 65) and in small amounts.
However, acrylamide formation is not intense in the mixture of asparagine and
ascorbic acid. When asparagine reacts with glucose, the acrylamide concentration is
83
Figure 64: Acrylamide formation in the mixtures of asparagine with ascorbic acid
Figure 65: Acrylamide formation in asparagine and ascorbic acid mixtures heated at250 °C and different times
84
usually 20 times higher.
We have also analysed asparagine and ascorbic acid anhydrous mixtures for 3-
aminopropionamide. At 170 °C 10 minutes heating was the first detection point for
this substance (Figure 66). At a molar ratio asparagine to vitamin C of 1:0.5 5 minutes
of heating time at higher than 170 °C we detected relatively high concentrations of
3-aminopropionamide: up to 350 µg/g. With increasing the heating temperature we
noticed a decrease in 3-aminopropionamide amounts formed.
Figure 66: 3-Aminopropionamide formation in the mixtures of asparagine with ascor-bic acid
7.8 Amadori compound
The Amadori compound 1-N -(asparaginyl)-5-azido-1,5-dideoxy-D-fructopyranose con-
sists of asparagine and monosaccharide residue. When we analysed it for acrylamide
85
and 3-aminopropionamide after heating, we detected, that both substances were
formed. However, the results show (Figure 67) that 3-aminopropionamide formed
in the first 2 minutes at 170, 190 and 210 °C was two times higher than acrylamide
formed. Moreover, with increasing the heating temperature the amount of acrylamide
and 3-aminopropionamide was increasing.
Figure 67: Acrylamide and 3-aminopropionamide formation from 1-N -(asparaginyl)-5-azido-1,5-dideoxy-D-fructopyranose
These results are in a strong agreement with other findings form model systems
where glucosylamine of asparagine is degrades to acrylamide via generated interme-
diate 3-aminopropionamide [39, 40, 41].
86
8 Conclusions
Within the scope of this dissertation extraction, clean-up and analytical methods for
analysis of acrylamide and 3-aminopropionamide were established.
As the simplest 3-aminopropionamide derivatized with dansyl chloride analysis
method the liquid chromatography with fluorescence detection was used. For acry-
lamide analysis in model systems a liquid chromatography with either UV or mass
spectroscopy was used. For the mass spectroscopy analysis method for acrylamide
derivatives with 2-mercaptobenzoic acid showed to be quite sensitive and rather ef-
ficient. Unfortunately, this method is not suitable for the coffee matrix because of
acrylamide low levels in it. For acrylamide measurements in coffee samples in routine
analysis the ion exchange chromatography with UV detection showed to be the most
efficient.
After the extraction, clean-up and analysis methods were established, the follow-
ing could be concluded� Arabica and Robusta coffee beans differ in acrylamide amounts formed. When
coffee was roasted in a laboratory roaster to common degrees for consumer, as
well as in the thermostatic oven under standard roasting conditions, Robusta
showed to have the highest amounts of acrylamide. It seems, that asparagine
is a limiting factor for acrylamide formation in coffee, because Robusta coffees
also contain higher amounts of asparagine than Arabicas.� The highest acrylamide amounts in coffee are formed at the very beginning
of the roasting process. After five minutes of roasting at temperatures higher
than 220 °C the amount of acrylamide is decreasing with increasing the roasting
time. Furthermore, acrylamide forms in lower amounts at higher temperatures
because of the faster elimination process.� The method for 3-aminopropionamide analysis was established. Unfortunately,
we could not detect 3-aminopropionamide neither in raw coffee beans nor in
87
roasted ones. In raw coffee beans the enzymatic conditions are unsatisfying
and probably enzymatic reactions do not take place. In the roasted coffee it is
difficult to detect 3-aminopropionamide because of possible low amounts of this
substance or its quite fast degradation into acrylamide. It also can be, that the
extraction method is not suitable for the coffee beans.
The experiments with model systems following conclusions could be made:� In the asparagine mixtures with glucose and sucrose it was observed that a
higher capacity of glucose form 3-aminopropionamide and acrylamide already
at the temperature of 130 °C, whereas in the mixtures with sucrose needs higher
temperatures to degrade into reactive compounds, acrylamide formation was ob-
served first only at 150 °C. Furthermore, the amounts of 3-aminopropionamide
and acrylamide formed were alike.� After some experiments were carried out with the model systems, we observed,
that asparagine mixtures with sucrose in the molar ratio 1:0.5 heated 170 °Cfor 7 minutes of heating produce the highest amounts of 3-aminopropionamide,
whereas optimal conditions for acrylamide formation were asparagine and glu-
cose mixture 1:0.5 heated at 250 °C for 1 minute.� After pure asparagine was heated at temperatures >200 °C, surprisingly acry-
lamide formation was detected. However, the amounts of this neurotoxin were
extremely low. Heated asparagine to 170 °C performs fast intramolecular cy-
clization forming maleimide and so prevents the formation of acrylamide.� As we expected in the asparagine mixtures with ascorbic acid heated at quite
high temperatures 3-aminopropionamide and acrylamide were detected. It was
noticeable, that for this reaction in order to form acrylamide and its precursor
higher temperatures and longer heating time are needed.� Heating of the Amadori compound 1-N -(asparaginyl)-5-azido-1,5-dideoxy-D-
fructopyranose showed that it was able to form 3-aminopropionamide and acry-
88
lamide. This is in agreement with other studies where it was said that as-
paragine is used as a skeleton for acrylamide’s formation.
89
9 Bibliography
References
[1] Material Safety Data Sheet: Safety data on Acrylamide. The Phys-
ical and Theoretical Chemistry Laboratory Oxford University. URL:
of coffee brews in relation to the roasting degree. LWT 30, 292-297.
98
[80] Nebesny E., Budryn G., 2003, Antioxidative activity of green and roasted coffee
beans as influenced by convection and microwave roasting methods and content
of certain compounds. Eur. Food Res. Technol. 217, 157-163.
[81] van der Stegen G.H.D., Essens P.J.M., van der Lijn J., 2001, Effect of roasting
conditions on reduction of ochratoxin A in coffee. J. Agric. Food Chem. 49,
4713-4716.
[82] Romani S., Pinnavaia G.G., Dalla Rosa M., 2003, Influence of roasting levels
on ochratoxin A content in coffee. J. Agric. Food Chem. 51, 5168-5171.
[83] Casal S., Mendes E., Alves M.R., Alves R.C., Beatriz M., Oliveira P.P., Ferreira
M.A., 2004, Free and conjugated biogenic amines in green and roasted coffee
beans. J. Agric. Food Chem. 52, 6188-6192.
[84] Moret S., Smela D., Populin T., Conte L.S., 2005, A survey on free biogenic
amine content of fresh and preserved vegetales. Food Chem. 89, 355-361.
[85] Bartzatt R., 2003, Dansylation of aromatic, aliphatic, and medicinal carboxylic
acid compounds in 1 M Na2CO3 buffer. Anal. Chim. Acta 488, 203-209.
[86] Naassner M., Mergler M., Wolf K., Schuphan I., 2002, Determination of the
xenoestrogens 4-nonylphenol by high-performance liquid chromatography and
fluorescence detection after derivatisation with dansyl chloride. J. Chromatogr.
A 945, 133-138.
[87] Matta Aoun C., King C.C., 2004, Effects of Ascorbic Acid on French Fries Qual-
ity and Acrylamide content. 60th Southwest Regional Meeting of the American
Chemical Society, Fort Worth, TX, United States, 29.09-04.10, Conference,
Abstracts.
99
10 Curriculum Vitae
Personal information
Surname/First name: Bagdonaite KristinaDate of birth: 10.07.1979Place of birth: Radviliskis, LithuaniaNationality: LithuanianAddress: 4/7/39, Gaussgasse, Graz, A-8010, AustriaOffice: Institute for Food Chemistry and Technology, Graz Uni-