COMMENTARY

Molecular gastronomy
HERVÉ THIS is at the Institut National de la Recherche Agronomique (INRA), College de France, 11, Place Marcelin Berthelot, 75005 Paris.
e-mail: [email protected]

For centuries, cooks have been applying recipes without

looking for the mechanisms of the culinary transformations.
A scientific discipline that explores these changes from
raw ingredients to eating the final dish, is developing into its
own field, termed molecular gastronomy. Here, one of the
founders of the discipline discusses its aims and importance.

M
olecular gastronomy, having initially
become popular through international
workshops at the Ettore Majorana Centre
for Scientific Culture (Erice, Sicily),
is now spreading, with the creation
of professorships in France, Greece and Denmark,
seminars all over the world1, conferences, a European
programme2, PhD studies, and also recognition by
cooks of the interest of the technological applications
of this scientific discipline.
In September 2004, a commentary in Nature
Materials stated that “we can take our science into the
chef ’s domain — a developing area called molecular
gastronomy”3. This idea is not entirely accurate,
because this would make molecular gastronomy
(MG) only a technology, or an application of
chemistry and physics. However, MG is actually a
scientific discipline in itself, whose importance is
greater than simple technology transfer.
Initially, MG included modelling recipes, testing
old wives tales, inventing new dishes and introducing
new tools, methods and ingredients in the kitchen4.
However, as any intellectual enterprise is improved
when clearly defined, technological applications were
finally excluded from MG itself and a more precise Figure 1 This system is grouping microreactors in line and in parallel, in order to
definition was given in 2003: it was recognized that all materialize physical systems described by CDS formalism. Eight inlets can drive gases,
recipes are made of two parts that should be studied liquids and solids into channels where they are dispersed. The prototype has been
in their own right5. First the definition of the dish: designed in collaboration with the Institut für Mikromekanik Mainz (IMM), and constructed
for example, mayonnaise is obtained by dispersing oil by Volker Hessel and his team at IMM.
droplets in water, using surfactants from egg yolk to
cover the droplets and increase the stability (as any
emulsion, mayonnaise is not stable, only metastable);
pear jam is obtained by heating pears with sugar the dish, technical indications along with old wives
and water. The second part comprises detailed tales, proverbs, sayings and so on. We now call these
descriptions of the processes involved in preparing ‘culinary precisions’, or, for short, ‘precisions’.

nature materials | VOL 4 | JANUARY 2005 | www.nature.com/naturematerials 5

©2005 Nature Publishing Group

nmat1303.indd 5 9/12/04 12:22:18 pm

 COMMENTARY
formalism is obvious: formulas can describe systems
that were difficult to imagine, and processes can be
generalized. For example, one can keep the O/W +
G → (G+O)/W process, but change the ingredients:
whipping melted chocolate dispersed in water leads
to ‘Chantilly chocolate’, that is, a chocolate mousse
with no whipped cream nor whipped egg whites in
it; with melted foie gras dispersed in water, one can
produce ‘foie gras Chantilly’ and so on. A piece of
apparatus consisting of a series of microreactors has
been designed for the practical application of CDS
formalism (Fig. 1).
Another formalism has been added to describe
dishes and the organization of space9. The proposal
is to use symbols to describe particular dimensions
of objects and to arrange them according to
processes described by operators, thus allowing the
Figure 2 It has been written THE DEFINITION OF THE DISH description of an operating space for the creation of
by cooks that raspberries dishes. Such formalism is useful as it can lead to new
turn blue when they are Most dishes are composed of disperse systems, structures of dishes, with unexplored consistencies.
cooked in tin-covered pans. formally referred to as colloids. In their most basic
However, the experimental form, colloids consist of one phase (gas bubbles, oil STUDY OF PRECISIONS
test of this ‘culinary or water droplets, or solid particles) dispersed in a
precision’ refutes it. different continuous phase. Emulsions, gels, foams Since 1980, more than 20,000 culinary precisions
In fact, tin does not react and suspensions are all basic colloids, which have have been collected, mostly from French cookery
in the described conditions, many examples in the food world; for example, books1. Consideration of all the collected precisions
but some tin ions added milk is an oil-in-water emulsion, as is mayonnaise; shows that there is indeed a mixture of all
to raspberries do indeed whipped egg whites are foams and dry into solid possibilities: some precisions seem true and they
generate a dark blue colour foams (meringues); jams are physical gels; and are indeed true, some precisions seem true but they
because of the interactions cooked egg whites are chemical gels. Dishes are are wrong, some precisions seem wrong but they
with polyphenols generally more complex, however. For example, are true, some precisions seem wrong and they are
responsible for the colour ice cream has gas (air) dispersed (by foaming) in a wrong, and even some precisions do not lead to
of red fruits. Left: crushed condensed medium containing ice crystals, protein any judgment on their reliability. Intuitions about
raspberries with pure water. aggregates, sucrose crystals and fats (either crystals reliability are not worth considering — experimental
Right: crushed raspberries or liquid droplets) — a difficult material to give a tests have to be done (see Fig. 2).
with Sn2+ ions. technical name to! For example, chefs cook green beans in boiling
Disperse systems can be described by a salty water; when the beans are cooked (decided either
‘formalism for complex disperse systems’ (CDS by smell, texture or flavour), they are strained, then
formalism) based on that invented in 1791 by immediately refreshed in icy cold water in order to,
Antoine-Laurent de Lavoisier6 for chemistry. chefs say, “fix the chlorophyll”10. When asked, chefs
By using the following abbreviations: G (for gas); admit that “fixing the chlorophyll” means keeping
O (oil, that is, any liquid fat, possibly with odorant the brilliant green colour of vegetables — but they
molecules dissolved in it, such as olive oil, but also forget that raw green beans are actually less green
melted foie gras, melted cocoa butter…); than when they are cooked! Is icy cold water useful
W (water, that is, any aqueous solution) and to keep the green colour? Having observed that cold
S1, S2, S3… (solids of any kind); and with operators water has no effect on the visual aspect of green beans,
such as the following: / (dispersed into), + (mixed in ordinary culinary conditions (of course, over-
with) and so on, formulas can be constructed to cooked beans turn yellow-brown), we investigated the
describe physical systems. matter further. Chlorophyll pheophytinization (the
For example, the process of whipping cream replacement of the magnesium atom by a proton, at
can be written: O/W + G → (G+O)/W. Equally this the centre of the chlorophyll molecule) changes the
formalism can be used to describe new physical green colour of chlorophyll a into an olive–brown
systems — for example ((G+O+S1)/W)/S2 refers to colour11, and it has been shown that the correlation
a dish made of a gel that contains a water solution between chlorophyll pheophytinization and the green
where air bubbles, solid particles and oil droplets colour (as measured by colorimetry) of cooked beans
are dispersed; I called it a Faraday in honour of the is poor. By UV spectroscopy of extracted pigments, it
great experimentalist and submitted the recipe for a has been found that cooling the green beans in ice has
‘Faraday of scallops’ (served in 2003 at the Restaurant no effect on the colour (C. Gremillet and J. Martin,
Pierre Gagnaire, Paris) to the Royal Institution7,8. manuscript in preparation).
It is interesting to observe that the CDS The case of mayonnaise helps to understand the
formalism can describe globally any complex reason why culinary precisions arose. Most modern
disperse systems, instead on focusing on the interface recipes say to add oil to a mixture of egg yolk and
only. (Note that it would be easy to add new vinegar (definition), but many authors add precisions.
operators as necessary to improve the description Some state that the temperature of the room where
of a particular system.) The usefulness of this the mayonnaise is made should not be too hot or

6 nature materials | VOL 4 | JANUARY 2005 | www.nature.com/naturematerials

©2005 Nature Publishing Group

nmat1303.indd 6 9/12/04 12:22:21 pm

 COMMENTARY
that the sauce should be made in a vessel lying on ice
cubes; others state, on the contrary, that a cold room Box1: NO NUTRITIVE PILLS — EVOLUTION PREVENTS IT
temperature is responsible for failure; and others state
that oil and eggs should be at the same temperature.
Other causes of failure are frequently given: for A prediction made in 1894 by the French chemist Marcelin Berthelot was that in
example, many cooks, in particular in France, the year 2000, humankind would have dropped agriculture and cooking would
think that mayonnaise fails when it is made by eat nutritive pills16. The year 2000 is over, yet we still stick to our cassoulet,
women having their periods, or that the cook should roasted chickens and French fries. Why was Berthelot’s prediction so inaccurate?
always turn the spoon or whisk in the vessel in the Berthelot was very enthusiastic about science in general and chemistry in
same direction. particular, but he forgot that our food behaviour is dictated by both biological
All these precisions were tested. It was easily needs and our culture. First of all, our many sensory receptors (for odour, taste,
checked that women’s periods and the direction of consistency, temperature, pain …) have evolved so that, as Brillat-Savarin wrote,
whipping do not lead to failure. The question of “The Creator, by forcing humankind to eat for his living, invited it by appetite
temperature was also checked, with eggs from the and rewarded it by pleasure”17. From the point of view of receptor stimulation,
fridge (4 °C) and oil at room temperature (35 °C), pills are much weaker than cassoulets or sauerkraut. Secondly, we eat mostly
or the reverse, and no effect of temperature was according to culture: Alsatians like Munster cheese, in spite of its very strong
observed. This was expected, because mayonnaise odour but they do not eat durian fruits (a foul smelling but palatable fruit);
is ‘only’ an emulsion, that is, a dispersion of oil Asian populations like durian fruits but are repelled by Munster cheese.
droplets in water; the temperature range of the Finally the maximum energy content of food is in fat, but an easy calculation
emulsion stability is quite wide, and so temperature shows that to get the necessary energy for living, we should eat about 300 g of
becomes only the main factor for stability under the fat per day, which would require a lot of pills. Clearly, nutritive pills are still a
crystallization temperature of oil, or above protein fantasy — it is very unlikely that we shall eat them, even in the year 3000!
denaturation points (the first one is at 62 °C, for H. T.
gamma livetin)12.
However, some precisions were correct. For example,
it was sometimes written that oil should be added ‘drop
by drop’ at the beginning of the sauce preparation. or 15, but if oil were added only at 30 grams at References
And it is true that mayonnaise fails if too much oil is a time, then robustness would be reduced to 1. This, H. Reports of the INRA
Seminars on Molecular Gastronomy
added first, as water is then dispersed into oil, instead 15/30, or 0.5. More generally, for any recipe, many http://www.sfc.fr.
of oil into water. The latter is more stable because particular robustnesses can be calculated (time, mass, 2. http://www.inicon.net/
of the particular nature of the surfactants from the temperature) and these quantities can be compared 3. Donald, A. Nature Mater. 3, 579–581
(2004).
yolks: proteins and lecithins, which curve the oil/water with the number of corresponding precisions.
4. This, H. & Kurti, N. The Chemical
interface so that oil forms droplets. Intelligencer 65 (1995).
Discussion of mayonnaise could fill books, WHAT OF THE FUTURE? 5. This, H. Science des Aliments 23,
because the sauce is very popular and it has 187–198 (2003).
6. Lavoisier, A. L. Mémoires de
generated many culinary precisions. The question There is a fear that the advancement of science l’Académie des Sciences 492 (Paris,
is why do we have this mixture of right and wrong would mean the development of nutrition pills and 1782).
ideas about something as popular as mayonnaise? the abandonment of foodstuffs as we know them 7. This, H. Supplement to the British
Consideration of collected precisions led to an (Box 1). On the contrary, MG science proposes Nutrition Journal (in the press).
8. This, H. http://www.pierre-gagnaire.
assumption: do precisions arise when recipes can fail? a wealth of new dishes and processes13, and thus com (‘Science and cooking’ section).
To test it scientifically, a mathematical description has much to contribute to the culinary art. This is 9. This, H. Pour la Science 322, 6 (2004).
of how easily recipes fail was needed. This is why we why the European project Inicon2, creating some 10. De Gencé Encyclopédie de la Vie
Pratique 638 (Librairie Nationale
introduced a quantity called ‘robustness’ of recipes, technology transfer, is centred on MG. This project
Des Beaux Arts, Paris).
with which we can now compare the effects of should contribute to the distribution of new results 11. Von Elbe, J. H., Huang. A. S., Attoe,
various parameters7. of MG to professional cooks, first, and later to all the E. L. & Nank, W. K. J. Agric. Food
For example, when mayonnaise is prepared, oil cooks at home. In France, where MG has been very Chem. 34, 51–54 (1986).
12. Eunice, L.-C. & Shuryo, N. Crit. Rev.
addition is an important parameter: if one litre of active for years, there are already many projects for Poultry Biol. 2, 21–58 (1989).
oil were added directly to one egg yolk (which presenting MG to the general public: TV shows and 13. This, H. The World of Food
contains about 15 grams of water), whipping would radio programmes, courses on MG, French Days of Ingredients 22–26 (June–July 2004).
generate the very unstable emulsion of water into oil, MG, experimental workshops in primary schools14, 14. http://crdp.ac-paris.fr/index.
htm?url=d_arts-culture/gout-intro.
and mayonnaise would fail. The maximum volume various projects in colleges15. All these programmes htm
for oil addition at each step is of the order of the could easily be done in other countries. 15. http://www.ac-paris.fr/article_sur.
volume of the egg yolk (15 g). If the volume of oil The main question of science is to look for php3?id_article=274
16. Berthelot, M. Science et Morale 119
added is lower than this limit, then mayonnaise mechanisms of natural phenomena. As Louis Pasteur
(Calmann-Lévy, Paris, 1897).
can succeed (as long as other parameters, such as said: “there are no applied sciences, but applications 17. Brillat-Savarin, J. A. La Physiologie
whipping efficiency, are not limiting factors). of science”. MG is a science, and its applications can du Goût 5 (A. Sautelet et Cie, Paris,
This is why ‘oil-addition robustness’ should depend be useful for the millions of people that cook daily. 1826).

also on the precision with which oil is added. If this We should not fear it, because instead of nutritive
precision is 1 gram, then oil-addition robustness pills, which are probably only a bad dream, we can
is defined as 15/1 (interval divided by precision), now envision a wealth of interesting new dishes.

nature materials | VOL 4 | JANUARY 2005 | www.nature.com/naturematerials 7

©2005 Nature Publishing Group

nmat1303.indd 7 9/12/04 12:22:22 pm