Mechanisms Determining Crispness and Its Retention in Foods with a Dry Crust - Food Colloids: Self-Assembly and Material Science - page 486

clear restrictions on the morphology and on the mechanical properties of the
solid matrix of crispy foods. 1,7 This will be worked out further below. In
general, for a product to be perceived by consumers as crispy, it must fulfil
certain requirements at the molecular, mesoscopic and macroscopic scales. 1
(Since the difference between 'crispness' and 'crunchiness' is not well estab-
lished, 1 we will only use the first term here.) We will discuss the mechanisms
acting at the various length-scales. The emphasis will be on the requirements for
crispy behaviour at the mesoscopic scale, as well as on processes at this length-
scale causing loss of crispness. But also the relevant aspects at molecular and
macroscopic scale will be discussed briefly.

34.2 Materials and Methods
Samples of dry biscuit (Knappertje, Verkade) and toasted rusk rolls (echte
beschuit, Bolletje) were used as models for dry crispy foods. Both these foods
are cellular solids. The products were bought in a local supermarket and stored
at

30%. For each experimental session a
new package was opened, and the first and last biscuits were discarded. Model
deep-fried snack crusts were prepared in a special mould in which a set amount
of batter could be deposited on a hydrophobic sieve (mesh size ¼ 200 mm)
giving a batter film of 1-2 mm thick above a silica gel with a water activity of
0.8. The whole was deep fried for 2.5 min in oil at 1801C. It resulted in flat
crispy products with a structure similar to the crust of deep-fried 'chicken
nuggets'. 8
Fracture experiments were performed using a Texture Analyser (Stable Micro
Systems TA.XT.plus). In one type of experiment the biscuits were fractured using
wedge penetration (wedge angle 301) at various speeds. In another type of
experiment, the fracture behaviour and acoustic emission were determined by
cutting the rusk rolls at a controlled speed in the range 0.2-40 mm s 1 using a
thin stainless steel razor blade fitted to the Texture Analyser. 9 In this way, at low
cutting speed ( o 1mms 1 ), the force drops and acoustic emission related to the
fracture of individual beams or lamellae forming the cellular sponge structure
could be recorded. Analogue data from the Texture Analyser were sent to a Bru¨ el
& Kjær Pulse front-end system, where they were converted into a digital signal
(65 kHz). All tests were done in an acoustically insulated room, and the motor of
the Texture Analyser was acoustically insulated.
Sound emission was recorded using a 1/2-inch, type 4189 free-field Deltatron
microphone (Bru¨ el & Kjær, Naerum Denmark) with a frequency band of
6.3 Hz to 20 kHz and a sensitivity of 50 mV Pa 1 . The analogue signal was
converted to a digital one using the same front-end system as for the force data.
Recording, replay and basic signal analysis were performed using Bru¨ el & Kjær
pulse Labshop software, and more detailed signal analysis using Bru ¨ el & Kjær
Sound Quality type 7698 software. Shortening of the time between two sound
events was done using Sound Forge (Sony Pictures Digital, USA).

B 231C under a relative humidity of B