8 thermal processing and food quality: analysis and...

8.1 Introduction: the importance of thermal processing

Most foods are submitted to thermal processes such as cooking, baking, roasting,extrusion cooking, pasteurisation or sterilisation. The first processes servingmostly to obtain particular sensory or texture features, the last two to assuremicrobiological safety and to eliminate some enzymatic activities that reducefood preservation.

The reactions that occur are of great importance for the production of aroma,taste and colour. From this point of view foods can be divided in two classes. Insome cases, these modifications must be reduced to a minimum because anatural and fresh appearance is required, as, for example, in drinkable milk orfruit juices. In other cases they are desirable, because they produce the specificsensory and texture features of foodstuffs, such as bread, cereals, chocolate,coffee, nuts, malt and cooked meat. However, when these reactions occur duringfood preservation, they always have a negative impact on quality.

One of the most important processes involved is the reaction of amino acids,peptides and proteins with reducing sugars and vitamin C, the process generallyknown as Maillard Reaction (MR) or non-enzymatic browning. This is a cascadeof complex competitive reactions and even if most of them are fundamentalorganic reactions, such as eliminations, aldol condensations, retroaldolfragmentations, oxidations and reductions, the fact that they occur simulta-neously and give rise to many reactive intermediates makes difficult theirinterpretation and control (Yaylayan 1997). For many years these transforma-tions have been studied for their technological impact, but more recently it hasbeen shown that they may be accompanied by a reduction in the nutritive valueof foods and by the formation of toxic compounds. The loss in nutritional quality

8

Thermal processing and food quality:analysis and controlA. Arnoldi, University of Milan

and potentially in safety is attributed to the destruction of essential amino acids,interaction with metal ions, decrease of digestibility, inhibition of proteolyticand glycolytic enzymes, and formation of anti-nutritional and toxic compounds.

8.2 The importance of the Maillard reaction

The first investigations on the MR were performed around 90 years ago byMaillard (1912). Subsequently, Amadori reported on the formation of arearranged stable product from glucose and amino acids which was named afterhim and Heyns reported a similar compound from fructose. An overall picture ofthe reactions involved was proposed by Hodge (1953). One of the most detaileddescriptions of the pathways that lead to the main Maillard Reaction Products(MRPs) can be found in a very complete review by Ledl and Schleicher (1990).Only some relevant points will be discussed here.

Reducing sugars are necessary for this reaction; generally they are amonosaccharide, glucose or fructose, or a disaccharide, maltose or lactose, andin some cases a pentose. Non-reducing disaccharides, such as sucrose, or boundsugars as in glycoproteins, glycolipids, and flavonoids, react only after hydrolysis,a process often facilitated by fermentation. The counterparts are free amino acidsor proteins. Fermentation is useful also to increase the concentration of free aminoacids, whereas in some cases (e.g. cheese) biogenic amines can react as aminocompounds. Small amounts of ammonia can be produced from amino acids duringthe Maillard reaction and large amounts are added for the preparation of aparticular kind of caramel colouring. The mechanism of this reaction has beenstudied very seldom in real foods, because it is too complex and the separation ofnon-volatile products is very difficult. On the contrary, most authors have usedsimple model systems in order to improve control of all the parameters and veryoften these results have been extrapolated to foods quite efficiently.

Consider now an aldose. The initial step is the condensation of the aldehydicgroup of the sugar with an amino group to give a relatively unstableglycosylamine 1 which undergoes a reversible rearrangement to give anaminoketose 2 (Amadori compounds) (Fig. 8.1). These intermediates have beenfully characterised in model systems and detected in many foods. The equivalentrearrangement of the fructose + amino acid adduct produces an amino aldosewhich is called Heins product. The reader will find a detailed description of theproperties and reactivity of the Amadori Heins compounds in Yaylayan andHuyggues-Despointes (1994).

At low water content and pH 3–6, they can be considered the main precursorsof reactive intermediates in model systems. Deoxydiketoses and deoxyaldoke-toses are degradation products of aminoketoses in the pH range 4–7. Ring openingfollowed by 1, 2 or 1, 3-enolisation are crucial steps in this transformation and arefollowed by dehydrations and fragmentations with the formation of many veryreactive dicarbonyl fragments. One of the most important end products of thereaction of 3-deoxyhexulose is 5-hydroxymethyl-furancarboxaldehyde 3 (HMF).

Thermal processing and food quality: analysis and control 139

In the presence of large amounts of primary amines, this compound is suppressedand replaced by pyrrole aldehydes of structure 4 as well as betaines 5 which havebeen demonstrated to derive from a common intermediate and not from HMF.Compounds 4 and 5 can be obtained also from maltose and lactose. A typicaldegradation product of 1-deoxydiketoses is furanone 6 which easily splits offformaldehyde to give furanone 7 which is the base for the formation of somerelevant low molecular weight coloured compounds. Furanone 7 is produced alsofrom the equivalent reaction of pentoses (Fig. 8.2).

Small C2, C3 and C4 fragments are produced from sugars by retroaldolcleavage (Fig. 8.3). Many of these compounds are very reactive and condense

Fig. 8.1 Amadori rearrangement.

140 Thermal technologies in food processing

readily with other retroaldol derivatives or sugar derivatives in which the carbonskeleton is still intact. The reaction with the amino group of amino acids resultsin the incorporation of nitrogen, a critical step for the formation of melanoidinsand important flavour compounds such as pyrazines.

One of the first observations by Maillard was related to the production of CO2

in the reaction mixture. This process is named Strecker degradation (Fig. 8.4).The mechanism involves the reaction of amino acids with �-dicarbonylcompounds to produce an azovinylogous beta-ketoacid that is decarboxylated.Thus the amino acid is converted in an aldehyde containing one less carbon than

Fig. 8.2 Relevant end products from 3-deoxydiketoses and 1-deoxydiketoses.

Fig. 8.3 Some of the compounds produced by retroaldol cleavage.


the original amino acid. Most of these aldehydes are very reactive intermediatesor have very peculiar sensory properties. In addition to the normal compounds,with cysteine the reaction yields also hydrogen sulfide, with methionine 2-methylthiopropanal and methanethiol. The Strecker degradation is thereforeresponsible for the incorporation of nitrogen and sulfur in many volatile andnon-volatile end products.

In recent years, very detailed experiments on the reaction between 13Clabelled monosaccharides or disaccharides (Tressl et al. 1995; Tressl andRewicki 1999) and amino acids have clarified the different Maillard pathways ofthese two important classes of food components, and a revised scheme for theMaillard reaction was proposed. In proteins the most relevant effect of theMaillard reaction is the non-enzymatic glycosylation which involves mostlylysine. The first glycation products are then converted to the Amadori productfructosyllysine that can form cross-links with adjacent proteins or with otheramino groups. The resulting polymeric aggregates are called advanced glycationend products (AGEs).

8.3 Thermal processing and food safety

The negative side of thermal processing is the possible formation of potentiallycarcinogenic MRPs, among which heterocyclic amines (HAs) have attracted agrowing interest during the past two decades. HAs include about twentydifferent derivatives that are primarily found at ppb level in cooked musclefoods (Sugimura et al. 1977). Many HAs have been shown to be carcinogenic inmice, rats, and non-human primates studies. However, epidemiologic studiesshowed conflicting data: some have shown an association between cooked meatand fish intake and cancer development and others no significant relationship

Fig. 8.4 Strecker degradation.


(Sugimura et al. 1993; Steineck et al. 1993). The International Agency forResearch on Cancer (IARC) regards some HAs as possibly or probablycarcinogenic to humans, and recommends minimising our exposure to them(IARC 1993). Minimising their formation requires knowledge of precursors,cooking conditions, reaction mechanisms and kinetics, so that the food industrycan choose optimal conditions for designing food processes and food processequipments.

HAs are found mostly in the crust of grilled and fried meat and fish and in thepan residue, but to a much lesser extent in the interior of meat (Skog et al. 1995).Cooking time and temperature are critical parameters in determining theiramount (Knize et al. 1994). The low concentration of HAs and the complexsample matrix of cooked foods make the analysis of these compounds verydifficult. Most data reported in literature refer to polar HAs in cooked musclefoods that are found at low ng/g level (for reviews see Skog 1993; Layton et al.1995). An improved method for the determination of non-polar HAs has beenpublished recently (Skog et al. 1998). A kinetics model for the formation ofpolar HAs in a meat model system has been proposed (Arvidson et al. 1998).

A detailed overview of the risk assessment can be found in a review byFriedman (1996). Data suggest that HAs are the only known animal coloncarcinogens that humans (except vegetarians) consume every day and, althoughvery difficult, it would be desirable to control their level in food (Fig. 8.5).

8.4 Thermal processing and nutritional quality

The glycation of protein by sugars affects negatively the nutritional value ofproteins (Friedman 1996). In fact, beside the destruction of particularly reactiveamino acids, such as lysine, all essential ones must be liberated from foodproteins by digestion and factors that decrease digestibility as cross-linkingreduces their bioavailability. Loss of nutritional quality has been demonstratedin heat treated casein, casein-glucose, and casein-starch and was explained withdecreased nitrogen digestibility, possibly also through impaired intestinalabsorption (Gumbmann et al. 1993). On the basis of labelled experiments,Mori and Nakatsuji (1977) affirmed that the reduction of the nutritive valuedepends on the reduction in intestinal absorption of the Maillard-induced lysinederivatives.

Besides the change in physiological properties of non-enzymatic browningproducts, there are other factors that could affect the nutritional quality, such asthe formation of toxic compounds during the heat treatment of food:heterocyclic amines (HAs), melanoidins and lysinoalanine (LAL). LAL is notexactly an MRP, but a cross-linking product which occurs in proteins treatedunder alkaline conditions owing to the fact that the �-amino group of lysine inthe protein chain reacts with dehydroalanine, derived from �-elimination ofgood leaving groups from cystine and O-substituted serine (Friedman 1999).The cross-linking reaction occurs mostly intramolecularly and results in the


Fig. 8.5 Relevant heterocyclic amines.

144T

hermal

technologiesin

foodprocessing

reduction of digestibility of the proteins involved. The loss of nutritional qualityis particularly relevant in infant nutrition where milk is the only source ofproteins and amino acids. Erbersdobler (1989) observed other adverse effects inexperiments with animals fed with thermally treated proteins, in particular,damage of the renal function, but these negative influences do not seem to besignificant in practical conditions.

Another important nutritional consequence of the MR in food is theformation of antioxidative materials. Chiu et al. (1991) carried out studies on theformation of antioxidants in the reaction between tryptophan and glucose orfructose. These MRPs have a positive effect in preventing lipid oxidation infoods and the scavenging of active oxygen species has been related to theseantioxidative effects. Chelation of metals is another important contribution.

There are also some scattered data on the formation of antibacterial materials(Einarsson et al. 1983, 1988). Products deriving from arginine/glucose andarginine/xylose seemed particularly efficient. However, the reported data referonly to model systems and it is not clear if these compounds really contribute tothe preservation of foods.

8.5 Thermal processing, food flavour and colour

The MR is responsible for most colour formation during thermal processing offood, e.g. bread baking, roasting of coffee and nuts, roasting of meat, kiln-dryingof malt. Although nobody can deny its importance in determining the quality offoods, the progress of knowledge in this field has been slowed by manymethodological difficulties, because most of the colour is due to high molecularweight polymers, which are named melanoidins. Some attempts have been madeto isolate melanoidins from foods, e.g. soy sauce (Hashiba 1973; Lee et al.1987), dark beer (Kuntcheva and Obretenov 1996), malt or roasted barley (Milicet al. 1975; Obretenov et al. 1991), coffee (Maier and Buttle 1973; Steinhart andPackert 1993), but their very complex, probably non-repetitive, structure haslimited their structural characterisation. Until now the only coloured compoundswhich have been fully characterised are some low molecular weight ones (Amesand Nursten 1989; Arnoldi et al. 1997; Ravagli et al. 1999; Hofmann 1998b,1998d; Tressl et al. 1998a, 1998b; Wondrack et al. 1997). Figure 8.6 shows thestructure of some coloured compounds isolated in recent years.

Generally foods contain proteins and more rarely free amino acids. In a recentwork (Hofmann 1998c), it was demonstrated that different compounds areproduced reacting glucose with glycine and alanine or �-casein. With free aminoacids the majority of coloured compounds were shown to have molecular weightbelow 1000 amu, whereas the reaction between glucose and casein leads to adrastic increase of the molecular weight. Hofmann (1998a) has proposed colourdilution analysis, a method that permits selection of the compounds of a mixturethat contribute mostly to colour. Another very important consequence of the MRis the formation of flavour, a complex sensory feature related both to taste and


smell sensations. From the viewpoint of taste, there is a generalised shift fromsweet to bitter sensations in part from the destruction of sugars and in part theformation of bitter compounds. The latter aspect has not been studied in detailyet, partly for methodological difficulties.

Much more can be said about aroma. Odorous compounds must be relativelysmall in size and lack polarity, therefore they can be studied by GC-MS muchmore easily than any other MRP. Hundreds of compounds have been identified.Nursten (1980–81) has suggested dividing them into three groups:

Fig. 8.6 Examples of low molecular weight coloured compounds recently reported inliterature: 19 (Ravagli et al. 1999); 20 (Hofmann 1998d): 21 (Hofmann 1998a); 22

(Hofmann 1998b); 23 (Tressl et al. 1998a); 24 (Tressl et al. 1998b).


1. simple sugar dehydration/fragmentation products, such as furans, pyrones,cyclopentenes, carbonyls, acids (Fig. 8.7);

2. simple amino acid degradation products, Strecker aldehydes and sulfurcompounds;

3. volatiles produced by other interactions: pyrroles, pyridines, imidazoles,pyrazines, oxazoles, thiazoles, aldol condensation products (Fig. 8.8).

Pyrazines give a major contribution to roasted and cooked aroma (Maga 1982;Arnoldi 1992); their organoleptic properties depend strictly on structure andquantitative structure-activity relationships (QSARs) have been proposed.

Flavours developed by using thermal treatment of selected precursors havebecome standard commercial flavours sold all over the world, and they aregenerally known in Europe as ‘process flavours’ (Manley 1994). They aregenerally sold as the final flavour, however microwave flavours may be sold asprecursor mixtures prior to the development of flavour.

Of course not only the absolute concentration, but also the typical note andthe odour thresholds of flavour are important in determining the sensoryproperties of a food. The reader will find many sensory data of MRPs in a verycomplete compilation by Fors (1983). Threshold values can sometimes berelatively high (ppm range) or extraordinarily low (ppb level or less), forexample 3-sec-butyl-2-methoxypyrazine can be perceived in water at 0.001 ppb,and 1,3,5-trithiane in water at 0.04 ppb.

Although MRPs are relevant impact compounds in foods, the presence of aodour-active volatile is not a direct measure of its importance to aroma. Specifictechniques have been developed to solve this problem: the most common are thegas-chromatography-olfactometry (GCO) technique (Cunningham et al. 1986)and the aroma extraction dilution analysis (AEDA) (Schieberle et al. 1990). The

Fig. 8.7 Relevant volatiles from sugar dehydration/fragmentation (Nursten 1980–81).


GCO bioassay characterises potent odour-active volatiles by sniffing the gaschromatograph effluent. Using a sniffing port, a trained operator notes thepresence of an odour at a particular retention index and records the sensorycharacteristics. In subsequent runs the sample is diluted by factors of three andthe analysis is repeated until the odour is no longer perceivable. Thecombination of these runs produces a corrected chromatogram with odourpotency defined as the area of the chromatographic peaks. Using this technique,for example, it was possible to determine the major odour potent compounds inglucose-proline model systems (Roberts and Acree 1994) which are burntcaramel, cotton candy and popcorn. They are diacetyl, 2-acetyl-1-pyrroline, 2-acetyl-1,4,5,6-tetrahydropyridine, 2-acetyl-3,4,5,6-tetrahydropyridine and fura-neol, and all of them are very small and apparently irrelevant peaks in the FIDchromatogram.

8.6 Maillard reaction and lipid oxidation

Another important reaction which occurs in processed and cooked foods andinfluences the aroma is lipid degradation. The intermediates of these reactionsare involved in the formation of food flavour. Experiments carried out with

Fig. 8.8 Relevant volatiles from other interactions (Nursten 1980–81).


model systems added with fat have demonstrated that several volatilecompounds are product by the interaction between MRPs and lipid oxidationproducts (Whitfield 1992). Most of them are heterocyclic compounds containingnitrogen or sulfur such as alkylthiazoles, long-chain alkylpyrazines (Chiu et al.1990; Huang et al. 1987), pyridines and thiophenes.

Phospholipids were added to model systems to mimic the formation of meataroma (Farmer and Mottram 1990; Farmer et al. 1989), and recently it wasdemonstrated that some edible oils can affect the pyrazine profile (Negroni et al.2000).

8.7 Controlling factors in the Maillard reaction

Understanding the factors that influence the MR is critical for achieving itscontrol. The formation of colour and flavour and the loss in nutritional value arethe aspects that have been studied in detail. The most important parameters thataffect the MR are: temperature, pH, water activity, and the structure of aminoacids and sugars involved. However, most of the information available has beencollected in model systems and not in real foods.

The effect of temperature is particularly strong and involves every aspect ofthe MR. For example, several experiments have shown that an increase intemperature and/or time of heating leads to an increase of colour developmentand aroma profile. Not only the amount of the MRPs is increased, but also theirnature is modified. For example it has been demonstrated that the carbon-to-nitrogen ratio, the degree of unsaturation and the chemical aromaticity of themelanoidins formed in model systems increase with temperature and time. Theformation of undesired compounds as HAs (see Section 8.3) is particularlysensitive to high temperature and a careful choice of the cooking conditionsallows them to be reduced to a minimum.

Considering flavour, at intermediate temperatures caramel or cooked flavoursare produced, while at higher temperatures the aroma profile is toasted orroasted. Nevertheless, the MR proceeds slowly also at room temperature andbrowning and off-flavour formation are responsible for the deterioration of foodduring storage.

The MR proceeds faster at low moisture level (Nursten 1980–81), and it isgenerally accepted that moisture values corresponding to a water activity around0.65–0.75 are the most favourable. The differences in colour and flavour of theouter part (where fast dehydration occurs) and wet inner part of baked or roastedfoods are a very clear proof of these effects. However, not all the MRPs aresensitive in the same way and studies on flavour have demonstrated thatdifferent classes of volatile components are more or less sensitive depending onwhether or not water is required for their formation. The complex network ofreactions that produce browning is far from being disclosed, but certainly manycondensations and dehydrations are involved and water activity is a criticalparameter of the reaction kinetics.


pH is another important parameter. Browning is much greater at pH above 7and the rate of colour formation can be reduced by decreasing pH. Flavour isalso influenced, for example pyrazines reductones and fission products arepreferred at high pH values, while furans, especially furfurals, prevail at lowerpH. Nevertheless, mixtures of compounds are formed at each pH. Carefulcontrol of the pH is necessary in model systems because of acid formation,whereas food pH is more stable because they have a buffered environment.

Another important parameter is the structure of the reactants: pentoses reactmore quickly than hexoses, that react faster than disaccharides, and aldoses reactfaster than ketoses. Non-reducing sugars such as sucrose, dextrines and boundsugars are involved only after hydrolysis. Aspartic and glutamic acids are relativelyunreactive, whereas lysine, the most reactive free amino acid, is reactive also inproteins owing to the �-amino group. The structure of the amino acids determinesthe formation of particular MRPs such as sulfur compounds from cysteine andmethionine, that, however, react only after hydrolysis or fermentation.

8.8 Methods of measurement

Browning is certainly a macroscopic effect of the MR and sometimes it isdirectly measured to determine the progress of the reaction. However, recentlymuch more sophisticated parameters have been suggested and used.

Non-enzymatic glycosylation of proteins is one of the most importantnutritional consequences and many methods have been developed for itsestimation. The first glycation product or Schiff base rearranges to a more stableketoamine or Amadori product. The Amadori product can form cross-linksbetween adjacent proteins or with other amino groups. The resulting polymericaggregates are called advanced glycation end products (AGEs).

During acid hydrolysis of proteins, the glucose derived Amadori productfructosyllysine reacts to give furosine in about 30% yields and a small amount ofpyridosine, whereas 50% reverts to lysine. Furosine was first detected byErbersdobler and Zucker (1966) in foods and it can be easily analysed by HPLCwith 280 nm detection. The corresponding reaction in milk and milk productsproduces lactulosyllysine that can be estimated by the furosine method, too. Thefurosine test has been used by several authors to study the progress of theMaillard reaction in different foods (Chiang 1983; Hartkopf and Erbersdobler1993, 1994; Henle et al. 1995; Resmini et al. 1990).

Carboxymethyllysine is another useful parameter. It was detected for the firsttime in milk by Buser and Erbersdobler (1986) and an oxidative mechanism wasproposed for its formation (Ahmed et al. 1986). Hewedy et al. (1991) havecompared several damage indicators for the classification of UHT milk, but haveshown that it is suitable only for monitoring very severe damage because it isformed only in small amounts.

Another useful parameter is 5-hydroxymethylfurancarboxaldehyde (HMF).Its formation in foods has been explained in two ways: via the Amadori products


through enolisation (in the presence of amino groups) and via lactoseisomerisation and degradation, known as the Lobry de Bruyn-Alberda vanEkenstein transformation (Ames 1992). Owing to this, recently it has beenproposed to measure separately the HMF formed only by the acidic degradationof Amadori products, called bound HMF and directly related to the MR, andtotal HMF related also to the degradation of other precursors (Morales et al.1997). This method is more reliable than the previous spectrophotometric one(Keeney and Bassette 1959).

�-Pyrrolelysine is another parameter that has been proposed to measure theMR in foods. It was observed for the first time in the reaction between glucoseand lysine (Nakayama et al. 1980). It is particularly useful in dry foods becauseit is very stable. Resmini and Pellegrino (1994) have proposed a method tomeasure protein-bound �-pyrrolelysine in dried pasta. The formation of thisMRP parallels very well the formation of furosine.

LAL, though not deriving from the MR, can be a useful parameter in cheesewhere the very low residual concentration of lactose impairs the MR and makesother decomposition routes not involving sugars more probable. Pellegrino et al.(1996) have proposed a very sensitive method to detect the addition ofcaseinates to mozzarella cheese based on LAL determination by HPLC withfluorimetric detection..

8.9 Application to the processing of particular foods

As indicated above, the MR is of critical importance for the production ofaroma, taste and colour of foods, that can be roughly divided into two classes.The first class contains foods where a natural and fresh appearance is requiredsuch as, for example, drinkable milk or fruit juices. In this case, the thermaltreatment is applied to sterilise and the modifications induced by the MR mustbe reduced to a minimum. Typical cooked, roasted or baked foods such as bread,cereals, chocolate, coffee, nuts, malt, and cooked meat belong to the secondclass. In this case, the MR is the tool to attain the specific sensory and texturefeatures. In this chapter three foods are discussed as examples of a very mildprocess (milk), a medium temperature process (bread), and a very drastic one(coffee).

8.9.1 MilkMilk is heated to improve its shelf-life and to kill disease-causing micro-organisms and viruses. Three processes are used:

1. pasteurisation at 85ºC for 2–3 s or at 72–75ºC for 15–30 s in a plate heater;2. UHT treatment in coils or plate at 136–138ºC for 5–8 s or by steam injection

at 140–145ºC for 2–4 s;3. autoclave sterilisation in closed bottles.


In each of them the time and temperature conditions must be selected carefullyto reduce the MR to a minimum.

Drinkable milk is probably one of the foods that has been more studied fromthis point of view and great importance is attributed to the preservation of itsnutritional value. Many parameters can be used to identify correctly the differentmilk classes and some of them are based on the MR. Although the pH is neutraland the sugar concentration is around 3.5%, the recent industrial technologies ofthe production of pasteurised and UHT milk are rather mild and milk containsonly lactose, a disaccharide relatively slow to react. In these conditions only thefirst steps of the MR occur and lysine glycation is the main consequence, whilearoma and colour remain practically unchanged. The main product is thereforelactulosyllysine, that can be quantified indirectly after acid hydrolysis asfurosine. Many methods have been proposed for its quantification, on one of thebest being proposed by Resmini et al. (1990). In raw and pasteurised milk, thelevel of furosine is around 3–5 mg/100 g protein, in UHT milk is 5–220 mg/100 mg, in sterilised milk is � 300 mg/100 g. Abnormal values indicate fraud,for example the addition of reconstituted milk powder.

Great interest has been dedicated to the nutritional value of dry milk powder,whose most important use is for the preparation of infant formulas. The mainprocesses used include preconcentration using film evaporating systems to 40–50% solids and spray drying. An alternative method is drum drying where theliquid is applied in a thin layer to a heated cylinder and after some time removedby a knife. With the latter method, however, the thermal exposure is muchhigher.

8.9.2 Baked productsWheat flour contains only very small amounts of free sugars and thefermentation process is very important to generate glucose and maltose thatare indispensable MR precursors. This can be improved by the addition of flourfrom malted grains very rich in �-amylase. Bread baking time and temperaturesvary very much as a function of the bread dimensions and constituents, ingeneral small products (about 45 g) may be heated for 18–20 min at 240–250ºC,while big products (1000 g) may be heated at 240–220ºC for 55–65 min.Naturally, especially in the case of large products, the heat transfer occursslowly and there is a temperature gradient from the outer to the inner part of thedough. Water evaporates efficiently only in the crust, while the inner partremains softer and colder. The extent of the MR and its consequences are nothomogeneous: the differences in colour, flavour and texture of the crumb andthe crust are probably part of the hedonistic pleasure of bread consuming. Incrumb the most important flavours are autoxidation products of linoleic acid,methional and diacetyl, while 2-methyl-3-ethylpyrazine, 2-acetylpyrazine, 2-acetyl-1-pyrroline and 5-methyl-(5H)cyclopenta(b)pyrazine and furaneol areresponsible for the caramel, malty and roasty notes of crust. Free proline is theprecursor of many volatile compounds that have been characterised in detail


(Tressl et al. 1985). The pleasant colour of crust is due in part to the MR and inpart to caramelisation of sugars.

8.9.3 CoffeeDuring roasting, coffee beans undergo many pyrolytic reactions, which lead tothe substances responsible for its particular colour and flavour. The knowledgeof green coffee composition is still incomplete and during roasting otherreactions take place besides the MR, so that the composition of roasted coffee,especially the non-volatile part, is still far from being clear. The process can bedivided into two phases: drying and roasting. An important parameter is heattransfer which is correlated with time and temperature of roasting. The intensityof colour is correlated with the final roasting temperature. The taste changesfrom slightly sweet (green coffee) to bitter.

Most of the constituents are transformed, with some exceptions such ascaffeine. Free amino acids and sucrose, which is around 8.0% in arabica greencoffee and 4.0% in robusta green coffee, disappear completely, but alsopolysaccharides are transformed, together with other compounds not directlyinvolved in the MR like chlorogenic acid. Any effect that changes theconcentration of the precursors in green beans before roasting influences thequality of the roasted material. At the end the condensation and caramelisationproducts are about 25%. More than 700 volatile compounds have been identifiedin coffee aroma: important classes deriving from the Maillard reaction arepyrans, pyrazines, pyridines, pyrones, pyrroles, thiazoles and thiophenes (Illyand Viani 1995).

8.10 Future trends

Extrusion cooking is a high-temperature/short-time technique was introduced in1930s in the cereal industries. Nowadays this process is used extensively in thefood industry for the production of expanded snacks, breakfast cereals, texturedsoy protein, and other foods. Due to the nature of this technique, involving hightemperatures, pressures and shear forces, the food matrix is subjected tochemical changes (gelatinisation of starch molecules, cross-linking of proteins)and the production of flavours can be induced by the Maillard reaction (Riha etal. 1996).

Another recent technique for food processing is microwave heating. Asexperienced also during domestic cooking, the aroma profile of foods cooked inthis way is a very critical point. The lack of volatiles is due to the speed ofheating and surface moisture and temperature; actually many aroma compoundsare steam volatile and during microwave heating they evaporate rapidly;moreover the surface aw, that generally is around 1.0, is very unfavourable forthe Maillard reaction. In order to enhance the flavour profile of microwavedfoods several approaches have been attempted, e.g. addition of commercial


flavourings or Maillard precursors, use of microwave absorbing susceptorsheets, combination of heating treatment (van Eijk 1992).

8.11 Sources of further information and advice

The reader will find a useful discussion of the consequences of thermaltreatment of the most relevant foods in Belitz and Grosch (1999). Every 3–4years the researchers meet for international symposia, whose proceedings are avery useful source of information (Waller and Feather 1983; Fujimaki et al.1986; Finot et al. 1990; Parliment et al. 1994; O’Brien et al. 1998).

8.12 References

AHMED M U, THORPE S R, BAYNES J W (1986), ‘Identification of N-�-carboxy-methyllysine as a degradation product of fructoselysine in glycatedprotein’. J. Biol. Chem., 61, 4889–94.

AMES J M (1992), ‘The Maillard reaction’. In: Hudson, B. J. F., Biochemistry inFood Proteins. Elsevier, London, 1992, 99–153.

AMES J M, NURSTEN H E (1989), ‘Recent advances in the chemistry of colouredcompounds formed during the Maillard reaction’. In: Lien, W. S., Foo, C.W. Trends in Food Science. Institute of Food Science and Technology,Singapore, 1989, 8–14.

ARDVISSON P, VAN BOEKEL M A J S, SKOG K, JAGERSTAD M (1998), ‘Kinetics offormation of polar heterocyclic amines in a meat model system’. J.Chromat. A., 227–33.

ARNOLDI A (1992), ‘Pyrazines. Part 1’. Riv. It. EPPOS 11, 25–31.ARNOLDI A (1993), ‘Pyrazines. Part 2’. Food Occurence Riv. It. EPPOS. 7, 33–9.ARNOLDI A, CORAIN E A, SCAGLIONI L, AMES J M (1997), ‘New colored compounds

from the Maillard reaction between xylose and lysine’. J. Agric. FoodChem., 45, 650–5.

BELITZ H-D, GROSCH W (1999), Food Chemistry, Springer, Berlin.BUSER W, ERBERSDOBLER H F (1986), ‘Carboxymethyllysine, a new compound of

heat damage in milk products’. Milchwissenschaft, 41, 780–5.CHIANG G H (1983), ‘A simple and rapid high-performance liquid-chromato-

graphy procedure for determination of furosine lysine-reducing sugarderivative’. J. Agric. Food Chem., 31, 1373–4.

CHIU E-M, KUO M-C, BRUECHERT L J, HO C-T (1990), ‘Substitution of pyrazines byaldehydes in model systems’. J. Agric. Food Chem., 38, 58–61.

CHIU W K, TANAKA M, NAGASHIMA Y, TAGUCHI T (1991), ‘Prevention of sardinelipid oxidation by antioxidative Maillard reaction products prepared fromfructose-tryptophan’. Nippon Suisan Gakaishi 57, 1773–81.

CUNNINGHAM D G, ACREE T E, BARNARD J, BUTTS R M, BRAELL P A (1986), ‘Charmanalysis of apple volatiles’. Food Chem. 19, 137–47.


EINARSSON H, GORAN S, SNYGG B G, ERIKSSON C (1983), ‘Inhibition of bacterialgrowth by Maillard reaction products’. J. Agric. Food Chem. 31, 1043–7.

EINARSSON H, EKLUND T, NES I F (1988), ‘Inhibitory mechanisms of Maillardreaction products’. Microbios, 53, 27–36.

ERBERSDOBLER H F (1989), ‘Protein reactions during food processing and storage– Their relevance to human nutrition’. Bibl. Nutr. Dieta, 43, 140–55.

ERBERSDOBLER H F, ZUCKER H (1966), ‘Untersuchungen zum gehalt an lysin undverfugbarem lysin in trockenmagermilch’. Milchwissenschaft, 21, 564–8.

FARMER L J, MOTTRAM D S (1990), ‘Interaction of lipid in the Maillard reactionsbetween cysteine and ribose: the effect of a triglyceride and threephospholipids on the volatile compounds’. J. Sci. Food Agric. 53, 505–25.

FARMER L J, MOTTRAM D S, WHITFIELD F B (1989), ‘Volatile compounds producedin Maillard reactions involving cysteine, ribose and phospholipid’. J. Sci.Food Agric., 49, 347–68.

FINOT P A, AESCHBACKER H U, HURRELL R F, LIARDON R (1990), The MaillardReaction In Food Processing, Human Nutrition and Physiology. Adv. LifeScience. Birkhauser, Basel.

FORS S (1983), ‘Sensory properties of volatile Maillard reaction products andrelated compounds. A literature review’. In: Waller G. R., Feather M.S.The Maillard Reaction in Food and Nutrition. ACS Symp. Ser. 215,American Chemical Society, Washington DC, 1983, 185–286.

FRIEDMAN M (1996), ‘Browning and its prevention: an overview’. J. Agric. FoodChem. 44, 631–53.

FRIEDMAN M (1999), ‘Chemistry, biochemistry, nutrition, and microbiology oflysinoalanine, lanthionine, histidinoalanine in food and other proteins’. J.Agric. Food Chem. 47, 1295–319.

FUJIMAKI M, NAMIKI M, KATO H (1986), Amino-carbonyl reactions in food andbiological systems. Development in Food Sciences 13., Elsevier,Amsterdam.

GUMBMANN M R, FRIEDMAN M, SMITH G A (1993), ‘The nutritional values anddigestibilities of heat damaged casein and casein-carbohydrates mixtures’.Nutr. Rep. Int. 355–61.

HARTKOPF J, ERBERSDOBLER H F (1993), ‘Stability of furosine during ion-exchange chromatography in comparison with reverse-phase HPLC. J.Chromatogr., 635, 151–4.

HARTKOPF J, ERBERSDOBLER H F (1994), ‘Model studies of conditions for theformation of N-�-carboxymethyllysine in food’. Z. Lebensm. Unters.Forsch., 198, 15–19.

HASHIBA H (1973), ‘Non enzymic browning of soy sauce. Use of ion exchangeresin to identify types of compounds involved in oxidative browning’.Agric. Biol. Chem., 37, 871–7.

HENLE T, ZEHTNER G, KLOSTERMEYER H (1995), ‘Fast and sensitive determinationof furosine in food’. Z. Lebensm. Unters. Forsch., 200, 235–7.

HEWEDY M, KIESNER C, MEISSNER K, HARTKOPF J, ERBERSDOBLER H F (1991),


‘Effect of UHT heating milk in an experimental plant on several indicatorsof heat treatment’. J. Dairy Res., 61, 304–9.

HODGE J E (1953), ‘Chemistry of browning reactions in model systems’. J. Agric.Food Chem., 1, 928–43.

HOFMANN T (1998a), ‘Characterization of the most intensely colouredcompounds from Maillard reactions of pentoses by application of colourdilution analysis’. Carbohydr. Res., 313, 203–13.

HOFMANN T (1998b), ‘Characterization of precursors and elucidation of thereaction pathway leading to a novel coloured 2H, 7H, 8aH-pyrano[2, 3-b]pyran-3-one from pentoses by quantitative studies and application of13C-labelling experiments’. Carbohydr. Res., 313, 215–24.

HOFMANN T (1998c), ‘Studies on the relationship between molecular weight andcolor potency of fractions obtained by thermal treatment of glucose/aminoacid and glucose/protein solutions by using ultracentrifugation and colordilution techniques’. J. Agric. Food Chem., 46, 3891–5.

HOFMANN T (1998d), ‘Identification of novel colored compounds containingpyrrole and pyrrolinone structures formed by Maillard Reactions ofpentoses and primary amino acids’. J. Agric. Food Chem., 46, 3902–11.

HUANG T C, BRUECHERT L J, HARTMAN T-C, ROSEN R T, HO C-T (1987), ‘Effect oflipids and carbohydrates on thermal generation of volatiles fromcommercial zein’. J. Agric. Food Chem., 35, 985–90.

IARC (1993), Monographs on the Evaluation of Carcinogenic Risk to Humans:Vol. 58. Some Naturally Occurring Aromatic Amines and Mycotoxins,International Agency for Research on Cancer, Lyon, 163–242.

ILLY A, VIANI R (1995), Espresso Coffee – The Chemistry of Quality. AcademicPress, London.

KEENEY M, BASSETTE R (1959), ‘Detection of intermediate compounds in the earlystages of browning reaction in milk products’. J. Dairy Sci., 42, 945–60.

KNIZE M G, CUNNINGHAM P L, AVILA J R, GRIFFIN E A JR, FELTON J S (1994)‘Formation of mutagenic activity from amino acids heated at cookingtemperatures’. Food Chem. Toxicol., 32, 55–60.

KUNTCHEVA M J, OBRETENOV T D (1996), ‘Isolation and characterization ofmelanoidins from beer’. Z. Lebensm. Untersch., 202, 238–43.

LAYTON D W, BOGEN K T, KNIZE M G, HATCH F T, JOHNSON V M, FELTON J S (1995),‘Cancer risk of heterocyclic amines in cooked foods; an analysis andimplications for research’. Carcinogenesis, 16, 39–52.

LEDL F, SCHLEICHER E (1990), ‘New aspects of the Maillard reaction in foods andin the human body’. Angew. Chem. Int. Ed., 29, 565–708.

LEE Y S, HOMMA S, AIDA K (1987), ‘Characterization of melanoidins in soy sauceand fish sauce by electrofocusing and high performance gel permeationchromatography’. J. Jap. Soc. Food Sci. Technol. 34, 313–19.

MAGA J A (1982), ‘Pyrazines in Flavour’. In: Morton, I. D., MacLeod, A. J., FoodFlavours. Part A. Introduction., Elsevier, Amsterdam, 283–323.

MAIER H G, BUTTLE H (1973), ‘Isolation and characterization of browncompounds of coffee’. Z. Lebensm. Untersch., 150, 331–4.


MAILLARD A C (1912), ‘Action des acides amines sur les sucres. Formation desmelanoidines par voie methodologique’. C. R. Acad. Sci., 154, 66–8.

MANLEY C H (1994), ‘Process flavour and precursors systems’. In: Parliment T.H.; Morello M. J.; McGorrin, R.J. Thermally Generated Flavors.Maillard, Microwave and Extrusion Processes. ACS Symp. Ser. 543,American Chemical Society, Washington DC, 16–25.

MILIC B L J, GRUJIC INJAC B, PILETIC M V, LAJSIC S, KOLAROV L J A (1975),‘Melanoidins and carbohydrates in roasted barley’. J. Agric. Food Chem.,23, 960–3.

MORALES F, ROMERO C, JIMENEZ-PEREZ S (1997), ‘Chromatographic determina-tion of bound hydroxymethylfurfural as an index of milk proteinglycosylation’ J. Agric. Food Chem., 45, 1570–3.

MORI B, NAKATSUJI H (1977), ‘Utilization in rats of 14C-L-lysine-labeled caseinbrowned by the amino-carbonyl reactions’. Agric. Biol. Chem., 41, 345–50.

NAKAYAMA T, HAYASE F, KATO H (1980), ‘Formation of �-(2-formyl-4-hydroxymethyl-pyrrol-1-yl)-L-norleucine in the Maillard reaction betweenD-glucose and L-lysine’. Agric. Biol. Chem., 44, 1201–2.

NEGRONI M, D’AGOSTINA, ARNOLDI A (2000), ‘Autoxidation in xylose/modelsystems’. J. Agric. Food. Chem., 48, 479–83.

NURSTEN H E (1980–81), ‘Recent developments in studies of the Maillardreaction’. Food Chem., 6, 263–77.

O’BRIEN J, NURSTEN H E, CRABBE M J C, AMES J M (1998), The Maillard Reactionin Foods and Medicine. Royal Society of Chemistry, Cambridge, 1998.

OBRETENOV T D, KUNTCHEVA M J, MANTCHEV S C, VALKOVA G D (1991), ‘Isolationand characterization of melanoidins from malt and malt roots’. J. FoodBiochem., 15, 279–94.

PARLIMENT T H, MORELLO M J, MCGORRIN R J (1994), Thermally GeneratedFlavors. Maillard, Microwave, and Extrusion Processes. ACS SymposiumSer. 543, Washington DC.

PELLEGRINO L, RESMINI P, DE NONI I, MASOTTI F (1996), ‘Sensitive determinationof lysinoalanine for distinguishing natural from imitation mozzarellacheese’. J. Dairy Sci., 79, 725–34.

RAVAGLI A, BOSCHIN G, SCAGLIONI L, ARNOLDI A (1999), ‘Reinvestigation of thereaction between 2-furancarboxyaldehyde and 4-hydroxy-5-methyl-3(2H)-furanone’. J. Agric. Food Chem., 47, 4962–9.

RESMINI P, PELLEGRINO L (1994), ‘Occurence of protein-bound lysylpyrrole-aldehyde in dried pasta’. Cereal Chem., 71, 254–62.

RESMINI P, PELLEGRINO L, BATELLI G (1990), ‘Accurate quantification of furosinein milk and dairy products by a direct HPLC method’. Ital. J. Food Sci., 3,173–83.

RIHA W E III, HO C-T (1996), ‘Formation of flavors during extrusion cooking’.Food Rev. Int., 12, 351–73.

ROBERTS D D, ACREE T (1994), ‘Gas chromatography – olfactometry of glucose-proline Maillard reaction products’. In: Parliment T. H.; Morello M. J.,McGorrin, R. J. Thermally Generated Flavors. Maillard, Microwave and


Extrusion Processes. ACS Symposium Ser. 543, American ChemicalSociety, Washington DC, 71–9.

SCHIEBERLE P, OFNER S, GROSCH W (1990), ‘Evaluation of potent odorants incucumbers (Cucumis sativus) and muskmelons (Cucumis melo) by aromaextract dilution analysis’. J. Food Sci., 55, 193–5.

SKOG K (1993), ‘Cooking procedures and food mutagens: a literature review’.Food Chem. Toxicol., 31, 655–75.

SKOG K, SOLYAKOV A, ARDVISSON P, JAGERSTAD M (1998), ‘Analysis of nonpolar heterocyclic amines in cooked foods and meat extracts using gaschromatography-mass spectrometry’. J. Chromat. A., 803, 227–33.

SKOG K, STEINECK G, AUGUSTSSON K, JAGERSTAD M (1995), ‘Effect of cookingtemperature on the formation of heterocyclic amines in fried meatproducts and pan residues’ Carcinogenesis, 16, 861–7.

STEINECK G, GERHARDSSON DE VERDIER M, OVERVIK E (1993), ‘Theepidemiological evidence concerning intake of mutagenic activity fromfried surface and the risk of cancer cannot justify preventive measures’.Eur. J. Cancer Prev., 2, 293–300.

STEINHART H, PACKERT A (1993), ‘Melanoidins in coffee. Separation andcharacterization by different chromatographic procedures’. Colloq. Sci.Int. Cafe [C. R.], 15th (vol. 2), 593–600.

SUGIMARA T, NAGAO M, KAWACHI T, HONDA M, YAHAGI T, SEINO Y, SATO S,MATSUKARA N, SHIRAI A, SAWAMURA M, MATSUMOTO H (1977), Mutagens– carcinogens in food, with special reference to highly mutagenicpyrolytic products in boiled foods’. In: Hiatt H. H., Watson J. D., WinstenJ. A., Origins of Human Cancer. Cold Spring Harbor Laboratory, 1561–77.

SUGIMARA T, WAKABAYASHI K, NAGAO M, ESUMI H (1993), ‘A new class ofcarcinogens: heterocyclic amines in cooked food’. In: Parke D. V.,Ioannides C., Walker R., Food, Nutrition and Chemical Toxicity. Smith-Gordon and Company, London, 259–76.

TRESSL R, REWICKI D, HELAK B, KAMPERSCHROER H (1985), ‘Formulation ofpyrrolidines and piperidines on heating L-proline with reducing sugars’. J.Agric. Food Chem., 33, 924–8.

TRESSL R, REWICKI D (1999), ‘Heat generated flavors and precursors’. In FlavorChemistry: 30 Years of Progress. ACS Symposiums Series, AmericanChemical Society, Washington DC, 305–25.

TRESSL R, NITTKA CH, KERSTEN E, REWICKI D (1995), ‘Formation of isoleucinespecific Maillard products from [1-13C]-D-glucose and [1-13C]-D-fructose’. J. Agric. Food Chem., 43, 1163–9.

TRESSL R, WONDRAK G T, KRUGER R P, REWICKI D (1998a), ‘New melanoidin-likeMaillard polymers from 2-deoxypentoses’. J. Agric. Food Chem., 46, 104–10.

TRESSL R, WONDRAK G T, KRUGER R P, REWICKI D, GARBE L-A (1998b), ‘Pentosesand hexoses as source of new melanoidin-like Maillard-polymers’. J.Agric. Food Chem., 46, 1756–76.


VAN EIJK T (1992), ‘Flavor and Flavorings in microwave foods: an Overview’.In: Parliment, T. H., Morello M. J., and McGorrin, R. J., ThermallyGenerated Flavors, ACS Symposiums Series 543, American ChemicalSociety, Washington DC, 395–404.

WALLER G R, FEATHER M S (1983), The Maillard Reaction in Foods andNutrition. ACS Symposium Ser. 215, Washington DC.

WHITFIELD F B (1992), ‘Volatiles from the interactions of the Maillard reactionand lipids’. Crit. Rev. Food Sci. Nutr., 31, 1–58.

WONDRAK G T, TRESSL R, REWICKI D (1997), ‘Maillard reaction of free andnucleic acid-bound 2-deoxy-D-ribose and D-ribose with �-amino acids’. J.Agric. Food Chem., 45, 321–7.

YAYLAYAN V A (1997), ‘Classification of the Maillard reaction: a conceptualapproach’. Trends Food Sci. Technol. 8, 13–18.

YAYLAYAN V A, HUYGGUES-DESPOINTES A (1994), ‘Chemistry of Amadorirearrangement products: analysis, synethesis, kinetics, reactions, andspectroscopic properties’. Crit. Rev. Food Sci. Nutr., 34, 321–8.


8 thermal processing and food quality: analysis and...

Documents