Molecular Gastronomy: Exploring the Science of Flavor (34 page)

methods for the culinary transformation of foods that contain sugars. Whereas

the Maillard reaction is a reaction of sugars with amino acids or proteins, cara-

melization involves only sugars. It is probable that the two reactions jointly

play a role in the cooking of most foods containing sugars, the share of each

depending on the relative quantities of sugars and proteins.

Although caramelization has influenced the taste and appearance of dishes

ever since sugars were first heated, exactly how these transformations take

place remains a mystery, and an economically important one at that: In France

alone the food processing industry produces 15,000 tons of caramel per year,

which are used in the making of milk, cookies, syrups, alcoholic beverages,

coffee, and soups.

| 227

A Scientic Tradition

The first scientific studies of caramel were done in 1838 by the French

chemist Étienne Péligot. For the next twenty years caramel was consigned to

purgatory, until M. A. Gélis, Charles Gerhardt, and Gerardus Johannes Mulder

proposed in 1858 to divide its nonvolatile component (making up 95% of the

caramelized product) into three parts: caramelan, caramelene, and caramelin.

Nonetheless, these substances, obtained from successive dissolutions with al-

cohol and water, were no more clearly defined, chemically speaking, than the

famous osmazome that Thenard and Brillat-Savarin claimed to constitute the

sapid principle of meats. None of the parts extracted by precipitation is consti-

tuted by a single type of molecule.

Investigation resumed in the early twentieth century. Caramel was then be-

lieved to contain humic acids, poorly understood reducing compounds whose

tanning properties are also found in lignite. The various compounds of the

volatile part of caramel were also discovered, including 5-hydroxymethyl-2-fur-

aldehyde and some twenty other compounds that contribute to its penetrat-

ing odor (including formaldehyde, acetaldehyde, methanol, ethyl lactate, and

maltol).

Subsequently it was observed that caramelan reacts with alcohols. Analysis

of the nonvolatile part nonetheless remained a nagging problem until 1989,

when modern research methods made it possible to detect the presence of a

derivative of glucose.

Water Eliminated

Sucrose is a disaccharide composed of glucose–fructose bonds. Each of

these two subunits has a skeleton composed of six carbon atoms. Five of these

atoms each carry a hydroxyl (–oh) group. The sixth one bears an oxygen atom

attached by a double bond, with a glycosidic bond such as –ch –o–ch – bind-

2

2

ing the two rings. Applying the same methods of analysis they had used in

studying the chemistry of sugars, the Grenoble researchers elucidated the main

features of the chemical transformations of the nonvolatile part of caramel.

Among other things they observed the formation of fructose dianhydrides, in

which two fructose rings are connected by two –ch –o– bonds, which in turn

2

define a third ring lying between them. Several molecules correspond to this

228 | investigations a nd mod el s

description because sugars come in many isomeric forms, which is to say that

molecules having the same atoms can differ if the atoms are linked by differ-

ent bonds.

Finally, the Grenoble chemists showed that during the caramelization of

sucrose, for example, the nonvolatile part results from an initial reaction dis-

sociating the sucrose into glucose and fructose. These elementary sugars then

recombine, forming oligosaccharides having various numbers of elementary

sugars: The glucose may combine with glucose or fructose, the fructose may

react with fructose, and so on.

These recent results are commercially important, for they make it possible

to consider polydextroses—used to give texture to dishes in which sugar is

replaced by intense sweeteners—as naturally occurring compounds. Because

polydextroses are naturally present in caramel, they are not subject to the same

system of regulation as other synthetic molecules. Moreover, the tendency of

various glucides to caramelize can now be investigated more easily.

Caramel
| 229

68

Bread and Crackers

The mechanical behavior of bread resembles that of plastic materials.

l e f t o u t i n t h e k i t c h e n , at room temperature, bread goes stale. Fro-

zen, it seems to change more slowly, but at what temperature must it be kept

in order to stay in the same state as when it comes out of the oven? 7°c (45°f)?

0°c (32°f)? –10°c (14°f)? Physical chemists at the École Nationale Supérieure

de Biologie Appliquée à la Nutrition et à l’Alimentation (ensbana) in Dijon

have sought to answer this question using their knowledge of polymers, which

are very long molecules formed by the linking of subunits called monomers.

This seemed to be a natural approach, for foods contain many polymers: The

molecules that constitute the starch granules in flour are linear or ramified

chains of glucose molecules known respectively as amylose and amylopectin,

proteins are chains of amino acids, and so on.

At high temperatures polymers are in a liquid state because they have suffi-

cient energy to move in a disordered fashion, allowing their mass to flow. When

polymers are cooled, they initially form a rubbery solid in which certain polymer

chains crystallize while preserving the ability to slide past one another. Then, at

temperatures lower than the temperature of vitreous transition, the chains are

immobilized and the material solidifies, with their crystalline parts dispersed in

an amorphous rigid part, or glass. The structure of the solid phase depends on

the cooling. When the cooling is rapid, the viscosity increases too quickly for the

molecules to be able to crystallize, and the vitreous part predominates.

230 |

Thus many foods are kinds of glass: Sugar cooked with water becomes

concentrated with the evaporation of the water and gradually forms a glass;

powdered milk, coffee, and fruit juice sometimes also appear in a vitreous

state. What about a fresh loaf of bread? Is it initially a rubbery solid that then

vitrifies or partially crystallizes as it goes stale? Martine Le Meste, Sylvie Davi-

dou, and Isabelle Fontanet at ensbana studied this question by recording the

mechanical behavior of various hydrate samples as a function of temperature

and comparing the reactions of loaves of bread with those of extruded flat

breads, such as crackers.

When one heats bread dough, which is essentially a mixture of flour and

water, the starch granules in the flour release their amylose molecules into

the water, as we have seen. As the bread cools, the amylose molecules form a

gel that traps the water and the amylopectin. In order to prepare variously hy-

drated breads, the Dijon team first completely dehydrated a series of samples

by placing them for a week in desiccators, where the water was absorbed by

phosphoric anhydride. The samples were then rehydrated under controlled

hygrometric conditions and coated with an impermeable silicone grease. A

viscoelastometer was used to measure the force transmitted by the samples

when they were deformed in a controlled way, yielding a coefficient of rigidity

known as Young’s modulus.

The researchers found that bread remains in a rubbery state as long as the

temperature is higher than the vitreous transition temperature, –20°c (–4°f).

On the other hand, analysis of the vitreous transition temperature as a func-

tion of water content showed that a part of the water does not freeze and that

it plays a plasticizing role.

Freezing Bread

These observations have practical implications. The many results obtained

by polymer chemists allow us to predict the changes in the mechanical proper-

ties of bread and its cousins as a function of their water content, crystallinity,

and so on. Among other things, even if the water that freezes is immobilized,

freezing will not arrest such changes as long as the temperature is higher

than the vitreous transition temperature. At temperatures between –20°c and

0°c (–4°f and 32°f), then, bread continues to undergo structural alteration. To

Bread and Crackers
| 231

preserve bread without compromising its textural characteristics, the freezing

temperature must be lower than the vitreous transition temperature.

The loss of freshness in bread had long been attributed to the phenomenon

of starch retrogradation, in which amylose progressively crystallizes, releasing

its water. The Dijon team observed instead a co-crystallization of amylose and

amylopectin into hydrated crystals. Lipids counteract the loss of freshness that

occurs over time because they bind with the amylose, forming crystals that

retard the co-crystallization of the amylose and amylopectin.

Nonetheless, the firmness associated with stale bread does not result solely

from this co-crystallization. The behavior of the amorphous, or vitreous, re-

gions seems to play a major role. Water is an important parameter in storage,

for it works to plasticize these regions, which in turn affects the rate and type

of crystallization that occur.

232 | investigations a nd mod el s

69

The of Alsace

The openness of the landscape is a crucial factor in winemaking.

t h e w o r l d o f w i n e a n d v i n e so little doubts the existence of differ-

ences in the overall natural environment—the
terroir,
as it is called—of wine-

growing regions that it has made them the basis for awarding protected desig-

nations of origin. Is this justified? Agronomists are accustomed to examining

how the particular features of a given viticultural site—its climate, soils, and

parent rock—affect the growth of its vines. Éric Lebon and his colleagues at

the Institut National de la Recherche Agronomique (inra) station in Colmar

have studied defined sections of the Alsatian landscape and shown that its

openness is at least as important as its capacity for retaining groundwater and

exposure to sunlight.

Wine growers seek to plant grapevines in conditions that favor the forma-

tion of berries, rather than leaves or branches, and the accumulation of sugars

(for fermentation) and aromas. Berries are able to ripen before the intemperate

weather of autumn causes them to rot only if the vines have begun to grow

early enough. For this reason it was long thought that sunshine was the chief

advantage of good
terroirs
.

At the request of the Centre Interprofessionel des Vins d’Alsace, the Colmar

agronomists continued research that had been begun in the 1970s in the Bor-

deaux region by Gérard Seguin and his colleagues at the Institut d’Œnologie

there. The team led by Seguin analyzed the importance of the soil, and the way

in which it nourishes the vine with water, in promoting the vine’s growth. The

| 233

best
terroirs,
it found, enjoyed a regular supply of water and periods of only

moderate drought, conditions that encourage the early ripening of the grapes.

Beginning in 1975, René Morlat and his colleagues at the inra station in

Angers studied the land in the legally protected red wine–producing vineyards

of the Loire Valley (Cabernet Franc from Saumur-Champigny, Chinon, and

Bourgueil). Their work confirmed the earlier observations and showed that

the more rapidly the soil heats up in the spring, the earlier the vine develops

and the more favorable the landscape is to successful cultivation. The Angers

agronomists suspected that the relevant climatic characteristics could be ana-

lyzed in terms of a landscape’s openness, determined by measuring the angle

of the horizon in relation to a horizontal axis for the eight principal directions

of the compass dial. Vines were supposed to develop differently in a basin,

where the degree of openness is low, than at the top of a hill, where the degree