short-barrelled single-screw extruder. It continued to be manufactured for many
years before the principles by which it was produced began to be understood and
indeed it was not until the 1970s and 1980s that the true nature of extrusion
cooking for direct expansion was determined by careful research.
8.5.1 The basic principles of direct snack extrusion
The transformation of starch-rich feedstocks such as maize grits, wheat, rice or
potato flour into hot melt fluids, which can be expanded as they emerge from a
die, occurs on the screw between the feed port and the die. In the early machine
this area was inaccessible to the scientist. At CCFRA in 1983
7
it became
possible to observe the changes in raw materials using a split barrel twin-screw
extruder. The first studies examined the development of wheat flour and maize
grits in the screws under the ideal running conditions to produce extruded
snacks. It was also possible to vary the processing parameters, including the
screw sections, and to follow the dynamic process variables and measure the
product characteristics. The process was examined on a standard screw system
comprised of three main sections, convey ing and mixing, back pumping and
shearing, pumping for extrusion at the die. This simple system was adopted after
studies with more complex screw designs were found to be no better and often
less stable.
1
Examination of a ‘dead-stop’ shutdown showed that in the conveying section
the screws were partly filled and the material was not compressed or sheared.
Almost all the heat was obtained from the barrel and the temperature rose to 70–
80ºC for a barrel profil e for three sections 2.5 L/D of 30/50/90ºC. The heat
transfer was related to particle size and smaller particles heated up more quickly.
The reversing or back pumping section caused the feedstock to fill some of
the conveying screws comple tely. This allowed the screws to pump the materia l
forward against the reversing section until the forward pres sure overcame their
resistance and set up a steady forward flow of material. The powder filling the
forward pumping screws was compressed to about 1 ml/g and filled the
reversing paddle elements. In the first paddle pair the pres sure was 20–30 bar
and the particles were sheared against each other and the metal elements. This
caused a large input of heat within a distance of a few mm and a rapid increase
in temperature. For the normal low moisture levels used in direct snack
extrusion of 16–18% moisture, the temperature rose from 80 to 150ºC in the first
paddle (12.5 mm along the shaft).
Examination of a maize or wheat flour feedstock showed that the crystalline
granules of the starch became amorphous and were squashed in the first pair of
paddles in the reversing section. The soft granules were deformed to flat
pancake shapes and then dispersed into a starch polymer continuum. Further
examination of the starch within the melted and sheared flour showed that the
starch polymers themselves had been degraded. Amylopectin (AM), which
occurs naturally as a very large branched polymer, was reduced from 10
8
D to
< 10
6
D, during a snack extrusion. The material within the reversing section was
174 Extrusion cooki ng