Extruded Foams: Production and Further Processing

Extruded Foams: Production and Further

Processing
W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

Institut für Kumststoffverarbeitung

Ponststr. 49
D-52062 Aachen
+49 241 8093806
[email protected]

1 INTRODUCTION
The main reasons for the increasing use of foamed plastics are, apart
from the saving of material, a number of advantageous properties such
as low weight, low thermal conductivity and good damping properties in
comparison with the solid material. As a rule, however, as the density
decreases, foaming leads to a disproportionate impairment of mechanical
properties. Microcellular foams represent an alternative to the familiar
foamed materials: on account of their extremely fine-celled structure they
have better mechanical properties at the same density. A method for the
production of microcellular foams with CO2 as the blowing agent during
extrusion is presented in Chapter 2.

Thermoforming has become very important for the production of

engineering mouldings with a three-dimensional structure from
thermoplastic semi-finished products in the form of film and sheet.
Because thermoforming, which competes with injection moulding, offers
advantages for thin-walled flat structures, one of the most important
applications is the production of cladding parts. Mouldings for motor-
vehicle interior trim occupy a special place in this context. On account of
their thin-walled structure, inside door panels, boot linings, hat-racks,
decorative parts in the area of the dashboard and centre console, and
inside roof linings are suitable for production by thermoforming. In
addition to the requirements on mechanical strength, as a rule parts for
interior trim of vehicles must satisfy high demands on aesthetic appearance
and feel, therefore most of the production methods employed comprise
several stages. Forming of a flat structure or a carrying component is
followed by operations of application of adhesive and coating. In this way,
for example, thermoplastic supporting parts reinforced with natural

Cellular Polymers, Vol. 21, No. 5, 2002 369

4Michaeli paper 369 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

fibres are coated with polyester fabric, PVC skin or textured TPO foam.
The various production processes are discussed in Section 3.

The properties of thermoplastic foams are essentially defined by the

cellular structure. To that extent, the quality of foams can be analysed by
examining the cellular structure. A possible way of determining the
structure of foams by means of digital image processing is described in
Chapter 4. Using the methods described there, an objective analysis of
foam structures is possible that is independent of place, time and
operator.

2 PRODUCTION OF MICROCELLULAR FOAMS WITH CO2

AS BLOWING AGENT

2.1 Properties of Microcellular Foams

Microcellular foams are characterised by an extremely fine-celled foam
structure with cells smaller than 10 µm and cell densities in the region of
109-1012 cells/cm3 in comparison with conventional foams with bubbles
larger than 100 µm and cell densities of 105-106 cells/cm3 (Figure 1)(1, 2).

This fine-celled foam structure leads to a lowering of the stress peaks in the
material and thus improves certain properties such as the impact resistance
and compressive strength of the foams in comparison with conventional
foams of the same density(3). If the cells are reduced in size down to the

Figure 1 Foam structure of polystyrene foams

370 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 370 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

range of the natural defects of the material, certain mechanical properties

can be improved up to the level of the solid material. Because the small cells
prevent crack propagation just like elastomer particles in impact-modified
polymers, for certain polymers the impact resistance can even be improved
above that of the unfoamed polymer(4).

Microcellular foams also have other interesting properties. On account of

the smallness of the cells, specially thin-walled products with a thickness of
less than 1 mm can be produced with a homogeneous foam structure.
Microcellular-foam specimens have a smooth, shiny surface. Depending
on the method of manufacture, this is achieved by means of a solid,
unfoamed boundary layer or by low surface roughness (Figure 2)(5).

Most of the previous attempts at extruding microcellular foams were

carried out with readily-foaming amorphous polymers, e.g. polystyrene
(PS). Building on these results, the aim of the investigations presented
here is to transfer the process to partially crystalline polymers such as
polypropylene (PP). As a first step, equipment is developed for producing
a temperature-controlled melt with a high proportion of blowing agent,
and a suitable process window is determined for the production of
microcellular foams from amorphous and partially crystalline
thermoplastics using simple filamentary dies. In the next step, experimental
tooling is constructed for extruding foamed sheets, taking into account
the rheological behaviour of the gassed polymer melt. The objective is the
laboratory-scale production of microcellular foamed PP sheets.

Figure 2 Surface quality of conventional and microcellular foamed sheets

Cellular Polymers, Vol. 21, No. 5, 2002 371

4Michaeli paper 371 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

2.2 Methods of Production of Microcellular Foams

In the production of microcellular foams, a distinction is made between
batch processes and extrusion processes(6). In the batch process, a
polymer sample is saturated with the blowing agent in an autoclave at
room temperature and elevated pressure. The rapid release of pressure
during opening of the autoclave causes the material to become
supersaturated with propellant gas, so that cell nucleation occurs. Cell
growth is then triggered by a defined temperature rise, for example in an
oil bath or hot-air stove. With the simple test set-up it was possible to carry
out numerous investigations for the production of microcellular foams by
the batch process. The use of this method is, however, limited mainly to
basic research, since the long loading times in the autoclave (> 24 hours,
depending on the material) make commercial application uneconomic.

In the extrusion process, as a first step a single-phase solution is prepared

from polymer melt and blowing agent. This is cooled to the required
foaming temperature and discharged through the extrusion die. As a
result of the pressure drop at the die outlet, just as in the batch process
the melt becomes supersaturated with blowing agent and cell formation
occurs. In conventional foam extrusion, the nucleation density is increased
by adding nucleating agents, which form defects in the melt, to which the
gas molecules become attached. This so-called heterogeneous nucleation
is described by(7):
-∆G*het
Nhet = Chet·f1·e kT (1)

where Nhet = rate of heterogeneous nucleation

Chet = number of heterogeneous nucleation sites in unit volume
f1 = frequency of attachment of gas molecules to the nucleus
∆G*het = energy of heterogeneous nucleation
k = Boltzmann constant
T = temperature, K

According to this, theoretically it is possible to increase the nucleation

density by increasing the concentration of nucleating agent, and therefore
achieve a homogeneous, fine-celled foam. In practice, however, there is
formation of agglomerations of nucleating agent, leading to the
development of individual large cells and therefore inhomogeneous foam
structures.

372 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 372 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

In order to achieve a high rate of nucleation, in the extrusion of

microcellular foams we therefore endeavour to obtain homogeneous
nucleation of the polymer/blowing agent mixture. This is described by:
-∆G*
Nhom = C0·f0·e kT (2)

where Nhom = rate of homogeneous nucleation

C0 = number of molecules of blowing agent in unit volume
f0 = frequency of attachment of gas molecules to the nucleus
∆G* = energy of homogeneous nucleation

The essential factor for increasing the rate of nucleation in this case is the
proportion of blowing agent, which is accordingly much higher than in
conventional foam extrusion. This requires correspondingly high process
pressures, to ensure that the blowing agent is dissolved completely. In the
production of conventional polystyrene foams, for example, the proportion
of CO2 used is up to 3% at die pressures in the region of 100 bar, whereas
in the production of microcellular foamed sheet 5-10% CO2 is used at
pressures above 200 bar(1).

To achieve a high cell density and a narrow distribution of cell sizes,

pressure relief for cell nucleation must take place in the shortest possible
time, as in the batch process. This can be explained by competition
between the nucleation of new cells and the growth of existing cells. At
a high pressure gradient, the melt becomes highly supersaturated,
causing nucleation of a large number of cells, which then grow through
inward diffusion of blowing agent from the surrounding melt. This causes
the concentration of blowing agent between the cells to decrease to the
extent that no new nuclei form, and the existing cells expand uniformly.
If, on the other hand, the pressure gradient is relatively small, initially
there is merely nucleation and growth of a few cell nuclei. Between these
cells, however, the concentration of blowing agent is still high enough for
the subsequent nucleation of additional cells. As these begin to grow at
a later point of time, an inhomogeneous foam structure forms, with a
wide distribution of cell sizes (Figure 3). To produce microcellular foams
with cell densities in the region of 109 cells/cm3 it is necessary to achieve
a pressure gradient in the region of 1 GPa/s (corresponding to
10,000 bar/s)(8). In comparison, pressure gradients of only approx.
500 bar/s are attained in conventional foam extrusion(9).

Cellular Polymers, Vol. 21, No. 5, 2002 373

4Michaeli paper 373 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

Figure 3 Effect of pressure gradient on cell nucleation

Owing to the good foamability of amorphous polymers, polystyrene (PS)

or impact-modified polystyrene (HIPS) in particular are suitable for
extrusion of microcellular foams(2). With these polymers, cell sizes of
10 µm at cell densities of 1010 cells/cm3 have been achieved at the
laboratory scale with throughputs of approx. 2 kg/h using hole-type dies.
In the manufacture of foamed sheet on an industrial scale, additional
forming of the melt with an annular die is necessary, and in this case cell
sizes in the region of 50 µm have so far represented the lower limit(1).

Process control in the production of microcellular foams from partially

crystalline polymers such as polypropylene (PP) is much more difficult
than with amorphous thermoplastics. On the one hand, CO2 is less
soluble in PP than in PS. Therefore even higher pressures are needed for
producing a homogeneous melt with a high proportion of dissolved
blowing agent. On the other hand, the melt strength of PP above the
crystallite melting temperature is so low that there is increasingly
coalescence of adjacent cells or bursting of individual bubbles. Both
processes lead to a lower cell density and inhomogeneous foam structures.
The main objective in conducting the tests is therefore to suppress these
effects by appropriate die design and process control.

374 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 374 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

2.3 Plant Engineering

The solubility of the blowing agent in the polymer melt increases with
increasing pressures and decreasing temperatures, therefore cooling of
the polymer melt after introduction of the blowing agent is required, in
addition to thorough mixing. In order to meet both requirements, in the
experimental plant the melt/blowing agent mixture is homogenised and
cooled using a static mixer. The static mixer is at the same time
constructed as a heat exchanger, with oil temperature control for
accurate adjustment of melt temperature. A melt pump is provided before
and after the heat exchanger, to control the throughput and the melt
pressure (Figure 4).

The mass throughput and therefore the die pressure are set by the second
pump. The rotation speed of the first pump determines the admission
pressure of the second pump and hence the required pressure in the static
mixer, independently of the throughput. In this way, pressures of up to
500 bar can be achieved in the mixer. The melt is supplied by a 60-mm
single-screw extruder, with its rotary speed controlled so that the
admission pressure of the first pump remains constant. The blowing
agent is metered in liquid form through a pressure transducer orifice into
the extruder barrel by means of a piston diaphragm pump. The prevailing
pressure in the extruder already causes some of the metered blowing
agent to go into solution. The undissolved blowing agent is dispersed so

Figure 4 Experimental set-up

Cellular Polymers, Vol. 21, No. 5, 2002 375

4Michaeli paper 375 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

finely, by numerous mixing elements on the screw, that the delivery

characteristics of the first melt pump are unaffected. This excess supply
of blowing agent is dissolved completely because of the high pressure and
melt cooling in the mixer.

For establishing the process window, a simple filamentary die with

diameters between 0.5 and 2 mm for extrusion of foamed extrudates is
used as the die. For producing foamed sheets, a film blowing head is used,
on account of the considerable expansion of the foams. As in the case of
conventional foam extrusion, the extruded sheet is pulled over a cooling
mandrel, cut on one side, laid flat and drawn off.

2.4 Test Procedure

For establishing the process window, tests were first carried out for
foaming of polystyrene (PS 158 K, BASF AG, Ludwigshafen) and a
standard polypropylene (Novolen 1102 K, Basell Polyolefins, Kehl) with
CO2 as blowing agent and with a filamentary die with a diameter of 1 mm.
The temperature of the static mixer and of the die was varied systematically
within a test programme, in order to determine the effect of temperature
variation on the structure of the foam. The mass throughput and the
proportion of blowing agent were kept constant. To ensure comparable
pressure conditions in the static mixer, the inlet pressure of the second
melt pump was also selected as constant. The pressure of admission into
the mixer is adjusted appropriately, in relation to the melt temperature.

The extrudates were foamed freely and cooled in the air. For determination
of foam density, several foam samples were measured and weighed at
each operating point. The foam structure of the samples was analysed on
the basis of SEM photographs of the sheet cross-section. The cell density
N was referred to the unfoamed material:
Nf
N= (3)
(1 - Vf)
with
4 πR3Nf
Vf =
3 (4)

where N = cell density

Nf = number of cells per cm3 of foam
R = mean bubble diameter
Vf = cell volume per cm3

376 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 376 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

2.5 Results of the Preliminary Tests

Figure 5 shows the effect of temperature variation on the foam structure
of the PS samples at a constant proportion of blowing agent of 5%. On
account of the pronounced plasticising action of CO2 in polystyrene, a
relatively wide temperature window is obtained between approx. 120°C
and 150°C melt temperature. The lower limit at 120°C is determined by
the maximum permissible pressure drop in the static mixer.

With the exception of the sample that was foamed at a melt temperature
of 152°C, all the samples have a uniform foam structure with a narrow
distribution of cell sizes. This homogeneous foam structure without the
use of nucleating agents shows that the proportion of blowing agent is
high enough to trigger homogeneous nucleation. As the melt temperature
decreases, the average cell size drops to 20 µm, because the higher melt
viscosity slows down cell growth and freezing-in of the sample occurs
earlier. For all test temperatures below 142°C the cell density is in the
region of 5·108 cells/cm3. This is an indication that the nucleation density
according to Eqation (2) is determined primarily by the concentration of
blowing agent. An influence of temperature on nucleation density could
not be demonstrated within the framework of these experiments.

For temperatures above 150°C, melt viscosity decreases to such an

extent that coalescence of cells is intensified and an irregular foam
structure with lower cell density is formed (Figure 5). Owing to the
pressure difference between two adjacent cells there is diffusion of the

Figure 5 Influence of melt temperature on foam structure (PS)

Cellular Polymers, Vol. 21, No. 5, 2002 377

4Michaeli paper 377 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

blowing agent from the smaller to the larger cell: the large bubble “eats”
the small one. To some extent there is also bursting of the low-viscosity
cell wall and hence collapse of some cells. This effect can also be
recognised in the expansion behaviour of the foamed extrudates (Figure 6).

At the low melt temperature there is considerable expansion of the

foamed extrudate, whose diameter then remains roughly constant owing
to the stable cell structure. An initial expansion can admittedly also be
observed when the melt temperatures are too high, but sample diameter
decreases again appreciably on account of the coalescing and collapsing
cells. In addition, the increased diffusion at higher temperatures also leads
to large losses of blowing agent on the surface(10).

Similar effects occur in the foaming of polypropylene. Once again, a

homogeneous foam structure forms at lower temperatures (Figure 7).
Compared with polystyrene, however, the processing window for
polypropylene is much smaller. Only a narrow temperature window is
available, between freezing-in of the material and a melt viscosity that is
too low. The sharp decrease in viscosity above the crystallite melting
temperature leads, at higher temperatures, to considerable coalescence
of cells and formation of an inhomogeneous structure (Figure 7).

Figure 6 Expansion of the PS samples

378 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 378 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Figure 7 Influence of melt temperature on foam structure (PP)

The solubility of CO2 in PP is also lower than in PS, so that only 3%

blowing agent can be dissolved at comparable pressures. This leads to a
lower cell density of 106 cells/cm3 at a cell size of about 100 µm.

2.6 Summing-up and Prospects

With the aid of the plant design described above, consisting of a static
mixer with a melt pump installed before and after it, the pressure and
temperature of the melt can be set over a wide range independently of
the throughput, and it is therefore possible to dissolve the high proportions
of blowing agent that are needed for the extrusion of microcellular foams.
The extruded polystyrene foams have cell densities of 5·108 cells/cm3 at
a cell size of 20 µm and with a narrow distribution of cell size. Up to now,
the achievable cell density and cell size have been limited primarily by the
pressure drop in the static mixer. For a smaller pressure drop and hence
a higher pressure level at mixer outlet, higher proportions of blowing
agent and moreover higher cell densities can be achieved. Therefore a
new design of the static mixer is now being developed and manufactured.

A particular aim of further research is to transfer the extrusion process

to an annular die, so as to be able to produce foamed sheets with
microcellular structure. In this case, especially for partially crystalline
polymers, process control and die design must ensure that there is stable
expansion of the foam, without coalescence or collapse of cells.

Cellular Polymers, Vol. 21, No. 5, 2002 379

4Michaeli paper 379 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

3 MAKING-UP OF POLYOLEFIN LAMINATES WITH LAYERS

OF FOAM AND FABRIC BY THE THERMOFORMING
PROCESS
Owing to their making-up from a variety of materials, many multilayer
lining components of the interior trim of today’s vehicles are expensive
to produce and must be classified as unfavourable from the standpoint of
recycling. The use of PVC mould skins is in decline especially in the USA,
as TPO is being used increasingly in current vehicle types(11). A trend
towards reducing the number of different types of materials for applications
in vehicle interiors is also seen in the European market. It can be assumed
that TPO and ABS, in particular, will become more important(12). Against
this background, there has also been increasing interest in the possibility
of using purely polyolefinic multilayer composites. Polyolefinic multilayer
composites are products that combine a layer of polyolefin foam with a
solid polypropylene backing (cf. Figure 8). Composites with a dimensionally
stable foam as the backing, which is coated with a softer foam or with
decorative fabric, are also possible (Figure 8). In addition to the two-layer
composites that have been described, composites constructed from three
layers are also feasible (Figure 8). The layers can comprise a solid backing
material, then a layer of foam, and then a surface fabric(13).

3.1 Applications of multilayered polyolefins

The choice of polyolefin semi-finished products for making multilayer
composites depends essentially on the requirements for the moulding

Figure 8 Structure of multilayered polyolefins

380 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 380 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

that is to be produced. For demanding interior applications the multilayer

composite must generally have a suitable decorative finish. Therefore a
fabric layer is a sensible choice for the visible surface of mouldings. If a
foamed material is used as the surface layer, its surface can be textured
to give it the appearance of real leather(14).

When using thermoforming for making mouldings for vehicle interior

trim, often a pleasant feel of the components (“soft-touch”) is required,
along with generally increasing quality requirements. For the dashboard,
some degree of impact protection for the occupants may also be
demanded. For inside door panels, surface fabrics tend to be used, with
relatively soft foams as a haptic interlayer, but impact protection in the
dashboard area can only be provided by higher-density foams that are
mechanically strong. The use of crosslinked polyolefin foams should be
mentioned here, as an alternative to the PVC foams that have been used
widely until now.

A positive effect that both fabric and foamed surface layers have in
common is the basically poor thermal conductivity. As a result, the
multilayer moulding has a heat-insulating effect, which can be utilised in
the design of heat-insulating components. The heat-insulating properties
also have a positive effect on the feel of the mouldings, because there is
little loss of body heat on contact, so that they are pleasantly warm to the
touch. With an excellent barrier effect against moisture, multilayer
mouldings with a layer of crosslinked foam are also suitable for watershields,
for example. These elements, used in vehicle doors, are increasingly
being demanded by vehicle manufacturers and often they must also
possess a damping action against external noise(15). Accordingly, fabric-
coated crosslinked foams are primarily of interest for this application.

In addition to the considerable potential for application of multilayer

polyolefin mouldings for lining components, another area of application
is materials handling(16). Transfer pallets, industrial packaging and transport
pallets have been produced with soft-touch surfaces by thermoforming.
In the case of transport pallets for delicate goods, a better distribution of
mechanical stresses can be achieved with a fabric or foamed surface layer.
This reduces the risk of breakage when stacking the pallets and during
transport. These advantages of fabric or foamed surface layers can also
be beneficial for transfer pallets. In particular, the scratching of polished
or otherwise delicate surfaces can be reduced if components are placed
in the depressions in the pallets.

Cellular Polymers, Vol. 21, No. 5, 2002 381

4Michaeli paper 381 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

An important possible application for multilayer mouldings of polyolefinic

materials is in connection with back-moulding technology. Composite
components with fabric and foamed layers, produced by thermoforming,
can be used as preforms for back-moulding. In the automotive area in
particular, there are good prospects for the thermoforming of multilayered
polyolefins as a constituent of two-stage processing. The production of
components in the two-stage process is then in competition with the
back-moulding of thick decorative materials of foam or fabric without
preforming. Owing to the high cost-pressure and high quality requirements
it is not always possible to use a single-stage process(17). Often the
available machines and especially the tooling cannot be used, and the
necessary decorative materials are too expensive. Back-moulding of
preforms is also appropriate in connection with the processing of PP-
fabrics having low resistance to thermal stresses. For conventional back-
moulding without preforming, mould cavity pressure of 50 to 100 bar,
depending on the fabric, must not be exceeded(18). When using preforms,
however, the backing material of the multilayer composite prevents direct
contact between the fabric and the melt. As well as the resulting decrease
in thermal stresses, mechanical stressing of the fabric by shear flows of
the melt are completely prevented by the backing material of the preform,
so that higher mould cavity pressure can be achieved.

3.2 Coating of Multilayer Polyolefin Semi-finished Products

In principle, various coating processes can be employed for the production
of multilayer polyolefin semi-finished products. Good bonding of the
layers can be achieved with adhesive, for example. However, during
subsequent processing of the multilayer composite by thermoforming,
the adhesive used is subjected to high thermal stresses. The effects on the
strength of the adhesive bond depend largely on the type of adhesive
used, and as the formulation of the adhesive is generally unknown, it
requires a lot of experience to be able to estimate these effects in advance.
Furthermore, the use of adhesives mainly affects the grade purity of
mouldings. However, one of the aims for the use of multilayer polyolefin
mouldings is better recycling in comparison with adhesive-coated mouldings
in modern vehicle interior trim, so that preference is given to adhesive-
free coating processes. Another argument for avoiding adhesives is the
demand by leading carmakers for less odour affecting the vehicle
occupants. Therefore there is a tendency to reduce the variety of plastics,
adhesives and other materials, so that the “cocktail” of odours and
emissions can be reduced(19). From this standpoint, therefore, prospects

382 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 382 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

are best for grade-pure multilayer polyolefins without adhesive films, for
which back-moulding with polypropylene can be used if required. The use
of recycled materials for back-moulding is also conceivable, and would
accordingly increase the environmental advantages of possible
components.

Flame coating can be considered as an alternative process for the

production of multilayer polyolefin semi-finished products without layers
of adhesive. With comparatively low process costs and high productivity,
this is a cost-effective alternative. Flame coating became important in the
production of sheeting with plastics components that can be attached by
surface melting, but especially in connection with the processing of PU
foams, for example as a component of carpets. Flame coating is in fact
already used occasionally for the coating of thermoplastic sheeting, but
comprehensive process know-how is not yet widely available.

Other coating processes that are available include coextrusion and

variants of “extrusion welding”. An advantageous aspect of both processes
is utilisation of the first heat for joining the layers. A disadvantage of both
processes is the comparatively high capital cost of equipment, which is
unacceptable especially for conducting basic research with frequent
changing of material and small production quantities. Both processes are,
however, very suitable for the coating of multilayer polyolefin semi-
finished products at high productivity.

In current research, thermally activated coating without adhesive is being

employed for making the required semi-finished products. The laboratory
scale coating station that is being used permits flexible production of small
batches with good reproducibility. The construction of the coating station
is shown schematically in Figure 9. Important elements of the station are
the system for heating with infrared radiation, which is equipped with
metal-foil radiating elements, and the rolling mill with adjustable roll gap.
The design of the equipment permits continuous coating of sheeting
semi-finished products. With radiation heating systems that can be
controlled independently, there is the option of heating one or both semi-
finished products. In addition, the intensity of radiation heating can
always be varied by controlling the power.

As an example of a multilayer semi-finished product produced with the

laboratory-scale laminator, Figure 10 shows part of a section photo of a
multilayer composite of PP backing material 1.2 mm thick, coated with

Cellular Polymers, Vol. 21, No. 5, 2002 383

4Michaeli paper 383 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

Figure 9 Laboratory-scale lamination device including IR radiators and

rolling mill

PP foam that is 2 mm thick. For production of this semi-finished product,

only the PP backing was heated with IR radiation. The adhesion between
the layers of the intermediate can therefore be attributed to wetting of the
foam surface with the melt pad on the surface of the heated backing
material. The width of the sheeting of multilayer composite produced is
about 190 mm, so the maximum roll width of the laminating station of
220 mm is almost fully utilised.

To determine possible processing windows in thermally activated

laminating without adhesive, operating points of the laminating plant
were selected at random. We investigated the production of a laminate
consisting of PP supporting material 1.2 mm thick, laminated with
crosslinked foam 2 mm thick. The test results are shown in Figure 11.
This relates lamination quality to the power setting of the radiating
element and the roll gap. Roll gap is shown as a relative dimension,
relating the distance between the rolls to the initial thickness of the foam
semi-finished product. Other parameters of the laboratory-scale laminator
– the feed speed (0.5 m/min) and the distance of the radiating element
(85 mm) – were kept constant for determination of the diagram.

Regarding lamination quality, in Figure 11 a distinction is made between

“weak” and “good” adherence. For this purpose the laminated composites
produced were submitted to manual testing, by pulling the crosslinked
foam from the PP backing at an angle of approx. 180°. The region of
“weak” adherence is characterised by a low resistance to the delamination

384 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 384 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Figure 10 Cross-section of PP backing laminated with PP foam

Figure 11 Laminate bonding depending on heating power and gap

this causes. In contrast, the region of “good” adherence was demarcated

on the basis that stripping of the crosslinked foam is only possible with
an increased expenditure of force, with formation of sharp radii at the
place of separation (cf. Figure 12). At the transition from the region of
“good” adherence to the region of “critical” heating power, stripping is
no longer possible without destruction of the crosslinked foam.

Cellular Polymers, Vol. 21, No. 5, 2002 385

4Michaeli paper 385 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

Figure 12 Manual testing of the laminate: “weak” and “good” adherence

It can be seen from the graph of lamination quality shown in the diagram
that good adherence of the crosslinked polyolefin foam on a PP backing
is obtained, with the heating equipment used, at heating power between
60% and 70%. With heating power below 60%, a melt pad does not form
on the PP sheeting, and accordingly adherence is inadequate. When the
heating power is set to above 70%, complete melting of the PP backing
causes breakage of the material, on account of the stresses applied to the
sheeting. In contrast to the heating power, the roll gap does not have any
notable immediate influence on adherence of the layers. When a
maximum gap of 0.9 is exceeded, the PP sheeting slips through, leading
to disturbance of material transport. It is not possible to go below the
critical gap of 0.05, owing to complete compaction of the foam.

Although adherence of the layers can be achieved within a large

processing window, the quantitative effects of heating power and roll gap
on the separation forces required for delamination of the composite have
not yet been investigated. In connection with the planning of further
series of tests, material tests according to DIN 53 357 are therefore
envisaged. The separation forces are measured using a standard tensile
tester equipped with the device illustrated in Figure 13. The roller ensures
that there is a constant pulling angle of 90° during determination of the
separation force. Morphological studies of foam quality are undertaken
using methods of electronic image processing. The OZELLA software
developed at the IKV will be used for this, as this is suitable for analysis
of closed-cell, dark-coloured foams(20).

386 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 386 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Figure 13 Device for testing laminate bonding according to DIN 53 357

3.3 Thermoforming of Laminated Polyolefins

3.3.1 Heating
Certain special characteristics must be taken into account when
thermoforming laminated polyolefins with a foam or textile layer. The
high heat-insulating effect of these layers means that efficient heating up
to the forming temperature can scarcely be achieved by conduction of
heat in the material(6). Therefore it makes sense to use heating systems
employing IR radiation. The great penetration depth of IR radiation,
especially with low-density foams, makes it possible for heat to be
transmitted directly to the thin cell membrane using IR-radiation heating(6).

In addition to the poor thermal conductivity of foam and textile layers, in

the thermoforming of laminated semi-finished products we must also
allow for the discontinuity of the thermal properties of the materials at the
layer boundaries. With a relatively large thickness of some layers, a
differentiated approach is required when examining heating of the
material. The radiation energy of an IR heating system is often largely
absorbed in the first layer(6). In addition there is reflection of radiation at

Cellular Polymers, Vol. 21, No. 5, 2002 387

4Michaeli paper 387 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

the layer boundaries, depending on the constitution of the layers.

Accordingly, heat transfer by radiation occurs disproportionately in the
layers near the surface, whereas in the interior of the material, heat
conduction mainly contributes to processes of temperature equalisation.
As heat conduction is a slow process compared with energy transfer by
radiation, there is a risk of thermal overloading in the surface layer of
laminated semi-finished products. To obtain a substantially homogeneous
temperature profile through the thickness of laminates it is necessary to
use IR radiation systems that heat from both sides. Depending on the
make-up of the laminate, two-sided radiation heating should be carried
out with different radiation power on the two surfaces of the intermediate.
A further increase in uniformity of the temperature profile over the
thickness of the semi-finished product can be achieved with surface
cooling using blowers. As a rule, relatively strong cooling of foam or
textile surfaces is required.

On the whole, to achieve a sufficiently uniform temperature profile

through the thickness of semi-finished products of laminated polyolefins,
a great many parameters must be set to the correct values. For
determining the settings required on the thermoforming machine used,
it is therefore advisable to use computer-aided heating simulation.
HEIZEN for Windows® was developed as a software tool for carrying out
these simulations. The simulation tool is based on a module for solving
the general differential equation of heat conduction using the Binder-
Schmitt difference method(21). The relations for absorption in the semi-
finished product are taken into account using the Lambert-Bouger law.

With consistent use of the simulation tool, it is possible to track down the
processing window of various polyolefinic laminates before carrying out
expensive series of tests. For the special case of an semi-finished product
with a thin textile surface layer, it is found in the heating simulation that
uniform heating of all layers of the composite is not possible with quartz
radiators by varying the available parameters. Checking of the results in
an experiment showed that for a 1.2 mm thick PP backing layer coated
with a 2 mm thick layer of crosslinked foam, heating is not possible with
the quartz radiators used and without a cooling blower, without damaging
a surface textile (PP spun fleece with specific weight of 70 g/m2).

3.3.2 The Forming Operation

During thermoforming of polyolefin laminates, process-dependent effects
lead to changes in the layer properties of the moulded article compared

388 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 388 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

with the properties of the unformed semi-finished product. As well as

changes in the morphology of foam and textile layers, alteration of the
adherence of the layers must also receive attention. It is to be expected
that the forming pressure used, in particular, will have a large influence
on the morphology of the moulded part.

For the largest possible scope for varying the forming pressure, we chose
a compressed-air sheet-forming machine for carrying out the forming tests.
In addition to pure compressed-air forming with a forming pressure of up
to 8 bar, this machine can also be used for vacuum forming and, steplessly,
all intermediate conditions including combined compressed-air/vacuum
forming. The construction of the machine is shown in Figure 14. The
arrangement of the forming station comprises the upper pressure bin, the
clamping frame and the lower mould. The conventional clamping frame is
only used for fixing the semi-finished product for the heating operation and
transport from the heating station to the forming station. The upper
pressure bin is provided for applying compressed air to the semi-finished
product and for sealing the semi-finished product against the tool. On the
mould table there is a mould box made of aluminium. Various moulds can
be fitted, within the set dimensions. As these moulds are not subjected to

Figure 14 Forming station with mould box and pressure box

Cellular Polymers, Vol. 21, No. 5, 2002 389

4Michaeli paper 389 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

any closing forces they can be made of epoxy resin, which can be filled with
aluminium grit. This means that the moulds can be made by casting, so
there is no need for expensive CNC machining.

Because it is possible to carry out investigations with various moulds, both

positive forming and negative forming can be employed. It is also possible
to use combined positive-negative moulds. In this way it also becomes
possible to investigate the influence of mould geometry on the resulting
adherence of the layers in the formed article. As an example, Figure 15
shows a mould inserted in the mould box.

Preliminary results regarding the effects on adherence of the layers in the

formed article were determined in tests with a polyolefin laminate
consisting of PE foam with a PE covering layer. Thermoforming of the
laminated semi-finished product led to localised delamination both on
sharp edges and radii, and in flat regions of the formed article. The
delamination that occurred on sharp edges and on narrow radii can be
explained by high degrees of local deformation, but in the case of local
delamination in flat zones – as shown in Figure 16 – we must conclude
that flaws are present. These defective areas presumably arise during
extrusion lamination, though perhaps also from surface defects of the
foam or as a result of external influences. Therefore studies are now in
progress for checking whether delamination can be prevented when
using thermal lamination.

Figure 15 Mould box with tool inserted

390 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 390 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Figure 16 Delamination at flaws

3.3.3 The Direct Lamination Process

For the production of formed articles in laminated polyolefins, consideration
may be given to the direct lamination process, as an alternative to the
processing of laminated semi-finished products(13,22). In this patented
variant of thermoforming, the bond between the layers in the formed
article is produced starting from semi-finished products that are supplied
separately and then subjected to forming together in the forming station.
An important advantage of the direct lamination process is therefore the
possibility of forming and bonding in one step(23). As well as the
associated elimination of an additional lamination process, this has the
further advantage of great flexibility. In direct lamination, different
combinations of semi-finished products can be processed from one cycle
to the next.

To achieve sufficient adherence of the layers in composites produced by

the direct lamination process, in some circumstances the coating product
must be heated or warmed, as well as the supporting semi-finished
product. Furthermore, the layers must be pressed together with sufficient
force. For gluing the layers, even comparatively low forming pressures
are generally sufficient, depending on the material and the forming
temperature. If regular welding of the layers is required, both material
layers must be melted at the surface. Moreover, shear flow must be
created in the contact area, so that intimate bonding of the two layers can
occur. This requires a squeezing pressure of at least 1 N/mm2 or a

Cellular Polymers, Vol. 21, No. 5, 2002 391

4Michaeli paper 391 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

corresponding forming pressure of 10 bar(24). At this level of pressure, in

connection with thermoforming we are already talking about High
Pressure Forming, and special moulds with locking devices must be used.
Therefore it is not possible to use conventional machinery for welding
layers in the direct lamination process.

A peculiarity of the direct lamination process that must be taken into

account is that air can be trapped at the boundary between the layers of
the formed article when the coating product is impermeable to air. An
example of this kind of defect in a formed article is illustrated in Figure 17,
which shows a flat, beaker-shaped form. The prominent trapped air in
the formed article on the right can be attributed to clamping both a solid
PP material and a crosslinked polyolefin foam that is also impermeable
to air, between the pressure bin and the mould of the compressed-air
forming machine. The same formed article without trapped air is shown
on the left. This result was achieved by thermoforming a laminated semi-
finished product consisting of 1.2 mm PP backing material coated with
crosslinked foam. An equally good result can also be achieved if the air
between the semi-finished products is removed immediately before direct
lamination. To accomplish this we can either apply a vacuum or the air
can be expelled as the semi-finished products pass through a rolling mill
together.

To avoid air trapping in the formed part – like that shown in Figure 17 –
ideally textile semi-finished products or open-cell foams should be used as

Figure 17 Two-layer formed part, with PP backing and TPO foam

392 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 392 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

the coating product in the direct lamination process. However, open-cell

foams are often unsuitable for soft-touch applications owing to their poor
mechanical properties and poor recovery capacity. Particle foams are
available as an alternative, and combine the positive properties of closed-
cell foams with the required porosity. In addition to materials specially
developed for direct lamination, particle foams are accordingly suitable as
foamed coating product for application in the direct lamination process.

If the direct lamination process is employed for making formed parts with
an air-permeable textile or foam surface layer, the choice of thermoforming
variant is limited. Owing to the air-permeability of the coating product,
the forming pressure and hence also the squeezing force for joining the
materials to the composite must be transmitted by the solid backing semi-
finished product. In the forming station, the backing material and coating
products should therefore always be arranged so that one or more air-
permeable semi-finished products are arranged between the mould and
the solid semi-finished product.

The direct lamination process can also be used for making three-layer
formed composites. As an example, Figure 18 shows the results of a
feasibility study. This shows a micrograph of the laminated composite with
solid PP backing, interposed haptic layer of PP foam and PP surface textile.
The PP needle fleece was formed by direct lamination on a precoated semi-
finished product consisting of PP backing and PP foam. The degree of
deformation of the slightly convex component is comparatively low.

Figure 18 Magnification of an all-PP three-layer laminate

Cellular Polymers, Vol. 21, No. 5, 2002 393

4Michaeli paper 393 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

3.4 Further Investigations of Forming

The thermoforming of laminated polyolefins is an interesting alternative
process for the manufacture of engineering forms. With proper utilisation
of the considerable application potential, future areas of application of
formed polyolefin composites are to be expected, mainly for vehicle
interior trim. Formed composites with just one grade of material, and the
possibility of making all-PP components with a decorative surface by
moulding of multilayer preforms offers prospects for increasing importance
of laminated composite forming, especially against the background of the
future EU guideline on used vehicles(25).

In current research, the processability of two-layer polyolefin composite

semi-finished products with PP backing material and crosslinked polyolefin
or PP foam layer and other composite semi-finished products with PP
foam and textile layers on a compressed-air sheet-forming machine is
being investigated systematically. The adherence between the layers of
the semi-finished products and formed parts is being analysed in
investigations of adherence based on DIN 53 357 for the testing of
plastics sheeting and film. Formed parts with laminated semi-finished
products are compared by examining the morphology of foam and textile
layers. In addition, the suitability of the direct lamination process for the
production of laminated polyolefin forms is being investigated.

4 ANALYSIS OF FOAM STRUCTURE BY DIGITAL IMAGE

PROCESSING
For many technical polymer foams, the structure of the foam is an
important characteristic for meeting the product requirements(26,27,28).
Analysis of this structure is necessary during optimisation of the production
process for achieving a stipulated foam structure (see Figure 19).

Until now, foam structure has mostly been analysed by manual

methods(29,30,31). This involves measuring and counting the cells. Owing
to the large number and generally abstract shape of the cells, this is very
time-consuming and can only be carried out selectively. Comparison of
the process variables with the foam quality achieved is expensive and
there is always a considerable time delay. Criteria for assessing foam
structure include cell size (average diameter), cell geometry (oval shape)
and number of cells in unit area. Unambiguous description of foam
structure is not possible with these manually-determined characteristics.

394 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 394 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Figure 19 Images of various foam structures

In (31), precisely defined features are determined from a table for

characterising foam by examining a sample of foam under the microscope.
Together with determination of density, a conclusion is drawn regarding
the degree of foaming of the sample. This method is strongly dependent
on the subjective impression of the observer. Newer methods of
measurement developed at the IKV permit analysis of foam structure by
digital image processing. In another approach(29), image processing is
employed for drawing the foam structure on a picture taken with a digital
camera. To some extent this is performed automatically by determining
a threshold value on the grey scale for demarcation between the light
areas of the cells and the dark matrix. Cells that are not identified by this
method are marked manually. Then the image is evaluated, determining
the number of cells in unit area and the cell size distribution. In this and
other methods(32,33,34), achievement of meaningful results requires
expensive sample preparation. Various software moulds for assessing the
cellular structure of polyurethanes are compared in(33). The considerable
scatter of the results reveals the inadequacy of these programs. In another
case(34), evaluation requires manual marking of the open-cell and closed-
cell regions. With the operations of preparation and marking, this
evaluation is very time-consuming. A mathematical approach for structural
analysis is presented in(35). In this approach as well, sample preparation
is very expensive and the calculation of the algorithm takes a long time.

Cellular Polymers, Vol. 21, No. 5, 2002 395

4Michaeli paper 395 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

A system for characterisation of foam structures of closed-cell foams is

described in(36). This system uses a commercial scanner for examining a
foam sample. Image evaluation therefore only takes place after a time
shift. Cell depth is not determined with this system. Only the cell
membranes are detected. Using this information, a binary image is
elaborated. A conclusion about the number of cells is based on the
proportion of black areas.

The coloration of the foams causes a problem in structure recognition.

The largest proportion is of a light coloration. Owing to this coloration,
suitable sensors and illumination systems must be used to attain the
necessary contrast for image evaluation. Impregnation of the upper layer
of cells with a gel or oil as coupling medium for light has been tried(29,37)
as a means of obtaining a high-contrast image. Impregnation caused
problems with adherence of the fluid to the foam samples. In addition, it
was not possible to ensure that only the cell membranes were impregnated
with the fluid. For these reasons, these imperfect preparation techniques
were again rejected. The time spent on preparation must also be taken
into account. For fast intervention in the production process, structural
analysis must be carried out as quickly as possible.

If a high-contrast image is available with a suitable digital imaging

technique, segmentation of the individual cells can take place. Methods
have been described(38,39) by which the edges of the cells can be
determined. These methods have the drawback of very long computation
time. For real-time processing, these methods need to be further
developed for optimum computation time.

4.1 Imaging Techniques for Analysing Foam Structures

This section describes various imaging techniques for structure analysis
of foamed polymers. Software for analysing dark-coloured, closed-cell
foam structures is presented in Section 4.1.1, and methods permitting a
complete 3D analysis of foams are presented in Sections 4.3 and 4.4.

4.1.1 The OZELLA Program (Optische Zell-Analyse / Optical Cell

Analysis)
The OZELLA program(38), which was developed at the IKV, determines
the course of the cell membranes on the surface of foam, using suitable
illumination and filtering of the data. On the basis of determination of the
membranes, the next step is statistical analysis of the number of areas

396 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 396 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

enclosed by the membranes. In this evaluation, information on the third

dimension of the cells is lost completely. The system can be used for
analysing closed-cell foams.

4.1.2 System Set-up

The arrangement of the system is shown in Figure 20.

The system consists of a CCD surface camera with macro lens and an
illumination unit, which illuminates the foam at a small angle (approx. 7°).
This form of illumination increases the contrast between cells and matrix
material. The cells show up clearly because of the shadows cast on the
cell walls. Irregularities in the foam can also be mistaken for cells. This can
be compensated by superimposing several images.

4.1.3 Methods of Analysis for Segmentation

Characterisation of foam structures requires segmentation of the cell
areas. Pre-processing of the image data is useful before commencing
determination of the cell areas. Pre-processing includes removal of image
interference (small-scale brightness fluctuations) by means of low-pass
filters (e.g. mean value filters) or by ranking methods (e.g. median filters).

Figure 20 The set-up of the OZELLA system for optical cell analysis

Cellular Polymers, Vol. 21, No. 5, 2002 397

4Michaeli paper 397 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

The threshold method, for example, can be used for segmentation. By

defining a limit of intensity, each pixel can be assigned to a membrane
or the interior of a cell. The limits of this method are shown very clearly
in Figure 21.

Only the cells with unambiguous shadowing can be detected. This is also
because of the grey-scale photographs of typical foam structures. They
do not exhibit bimodality either locally or globally. Therefore threshold
methods are only suitable for the detection of cell structures in certain
circumstances.

4.1.4 The Evaluation Steps in OZELLA

The evaluation steps in the calculation of cell areas can be understood
from Figure 22.

Firstly, filters are used for removing the brightness fluctuations from the
images obtained in the floodlight technique. This is followed by edge
detection. The cells can be detected on the basis of variations of
brightness between the membrane and the cell cavity. An algorithm that
is especially suitable for detecting diffuse edges has been described by Yu
and Mira(39). This algorithm has been implemented in OZELLA. In the
first step, the noise is reduced using a two-stage median filter. The cell
edges are determined with second-order two-dimensional gradient
formation. This is accomplished by combining the resulting median

Figure 21 Use of the threshold method for analysing a rigid (open-cell)

polyurethane foam

398 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 398 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Figure 22 Evaluation algorithm of OZELLA

image with a mean-value image(39). The gradient image is produced by

subtracting the mean-value image from the median image. Criteria for an
edge are, on the one hand, that the brightness gradient passes through
zero at the edge and, on the other hand, that the difference of two
adjacent median values is greater than zero. If further brightness variations
occur within the cells, these edges are detected as well. To filter out this
interference, an edge threshold value is determined, which is a measure
for the thickness of the edge.

The effects of changes of the parameters of the median filter, the mean-
value filter and the threshold value have been investigated(38), for reducing
interference in various photographic techniques. Suitable coefficients
were found for the investigation of dark-coloured foam samples. The
result according to edge determination is shown at top right in Figure 22.

To be able to determine a binary image with closed cell surfaces from this
image, the surfaces of a cell must be closed. For this purpose the edges
are completed and the interior regions of the cell are filled in relation to
a threshold value. Next, spuriously continuous regions must be separated.
A core expansion method was developed for this. Cells with very small
contact areas can be separated using this method. The result of an
evaluation is shown in Figure 23.

For evaluating the foam structure, the areas of the cells are plotted against
their frequency of occurrence (see Figure 24).

Cellular Polymers, Vol. 21, No. 5, 2002 399

4Michaeli paper 399 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

The OZELLA system can be used for analysing closed-cell structures.

When imaging an open-cell foam, the different regions of the actually
three-dimensional relief are projected in one plane. Evaluation using
OZELLA provides qualitative information on the structures of open-cell
foams. The methods presented in 4.2 to 4.3 are suitable for a more
thorough analysis.

Figure 23 OZELLA analysis of a sponge rubber seal (closed-cell)

Figure 24 Frequency distribution of cell size in a sponge rubber seal (closed-

cell)

400 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 400 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

4.2 3D Characterisation using the Stereo Imaging Technique

The degree of expansion of open-cell and closed-cell foams can be
investigated using three-dimensional analysis of foam structure. The
evaluation is to be employed on-line in foam production after an
experimental stage. Photographing with two cameras is required for
evaluation by the stereo image method. For simple adjustment and
investigation of different illumination and photographing angles, a test rig
was constructed that permits simple adjustment. The components of this
system are shown in Figure 25.

The system consists of two line cameras and a lighting set-up. The latter
uses a cold-light source with a converter for line illumination. This
converter has an adjustable cylinder lens for focusing. The foam sample
is secured in a holder on the traverse. The traverse moves the sample past
the cameras. In this way a continuous image is produced from the
individual lines of the cameras.

The surface must not be destroyed when the samples are cut. Furthermore,
a plane-parallel cutting plane must be achieved, so as to maintain a
constant distance from the camera sensors and hence a constant depth
of field. In the production of thermoplastic foams, fine cutters are used
for edge trimming or for cutting several strips. Trimming has a negligible

Figure 25 Set-up of a laboratory system for stereo image analysis

Cellular Polymers, Vol. 21, No. 5, 2002 401

4Michaeli paper 401 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

effect on foam structure(40). The method of sample preparation depends

on the imaging technique adopted, and on the properties of the foam
sample. For optical analysis it is necessary to produce a cut surface
through the foam. This cut surface must be representative of the foam
structure of the whole sample.

The proportion of light-coloured foams accounts for more than 80% of

the total output of foam(40). Accordingly, the system for analysing foam
structure must be capable of analysing these light-coloured foams as well.
In the systems used until now for analysis of foam structure(38), with light-
coloured foams there were considerable problems in obtaining an image
of high contrast, suitable for the analysis. Preliminary studies have shown,
however, that by using line cameras for imaging the foam sample,
together with optimised adjustment of the illumination time and intensity,
an optimum grey scale distribution can also be achieved with light-
coloured foams.

The first photographs with the arrangement shown in Figure 25 are

presented in Figure 26.

The scan angle was varied between the two measurements. Therefore
different shadow effects are obtained in the difference image. Until now
these images have been superimposed manually. For automated evaluation
of these images it is necessary to develop algorithms for calculating

Figure 26 Stereo photographs and their difference image

402 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 402 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

precise superposition of the images from the different viewing angles.

Differencing, without falsifying the measurement results, is only possible
after exact superposition.

With suitable algorithms, and knowing the imaging behaviour of the

camera, it is possible to reconstruct the 2D image information of the stereo
photographs as the 3D foam structure on the surface of the foam.
Algorithms with formation of a fundamental matrix(41) should be used for
the evaluation. At present these are used for calculating geometric relations
from images that were recorded from different directions of view.

4.3 Confocal Imaging for Reconstruction of 3D Structure

For confocal imaging, a small depth of field is set when imaging the foam
structures. Images are obtained with the camera at different distances
from the foam surface. Three photographs obtained in this way are
shown in Figure 27. With the different depths of field, in each case a sharp
image is obtained for one height structure of the foam.

From these images, a conclusion about the foam structure can be based
on calculation of the sharp regions. Using difference filters or an FFT
analysis, as used in(42) for determining correct focusing, the sharp regions

Figure 27 Confocal image acquisition with 3 planes

Cellular Polymers, Vol. 21, No. 5, 2002 403

4Michaeli paper 403 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

can be extracted from the individual images. If we now separate the sharp
regions from each plane and superimpose them, we get Figure 28. Cell
reconstruction can now be effected within the individual planes of the
image.

The reconstruction can be performed with closed-cell foams. In the case

of open-cell foam structures, determination of sharpness allows conclusions
to be drawn regarding the membranes that lie within the individual planes.
Research for this imaging technique is still in progress. The resolutions
and foam structures that can be investigated using this method are
currently being studied in a research project.

4.3 Summary and Outlook

Dark-coloured, closed-cell foams can be analysed using the OZELLA
system. By imaging with line cameras, with suitable lighting arrangements,
the contrast in images of light-coloured foams could be improved
considerably. The result obtained after evaluation is a frequency distribution
for cell size. The evaluation is yet to be extended to a choice of regions.
It is then possible to undertake evaluations of regions of integral foams.

Figure 28 Plane image constructed using the difference method

404 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 404 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

3D analysis is possible using stereo imaging and confocal imaging

methods. The algorithms for the evaluation are at present still in the
development phase. On-line imaging is also a possibility for the confocal
photographic technique. With this system, the line cameras are mounted
at different heights. This means that with the same settings they have
different focal planes. The extent to which this method is suitable for an
on-line investigation has yet to be investigated.

ACKNOWLEDGMENTS
The research presented in this report was undertaken with the financial
support of the Bundesministerium für Wirtschaft und Technologie
(BMWi) via the Arbeitsgemeinschaft industrieller Forschungsvereinigungen
e.V. (AIF) (“Untersuchung der Umformung und Verbindung von Polyolefin
- Mehrschichtverbunden”, No. 12513N, and “Strukturanalyse von
geschäumten Polymeren mittels digitaler Bildverarbeitung”, No. 12705N).

In addition, we are especially grateful to the companies Pixargus,

Aachen, and Krelus Infrared AG, Hirschthal, Switzerland, for the use of
experimental equipment. Special thanks are also due to Alveo AG,
Lucerne, Switzerland, the company Polymer-Tec, Bad Sobernheim, the
company TWE Technische Weberei, Emsdetten, Polymer Chemie GmbH,
Basell Polyolefine GmbH, Kehl and BASF AG, Ludwigshafen for actively
supporting the research by supplying test materials. We obtained valuable
encouragement and suggestions in discussions with the members of the
“Extrusion” technical advisory group and the member firms of the
Support Association of the IKV.

REFERENCES
1. M. Römer, Microfoam extrusion as a production process.
Thermoplastische Schäume - Innovative und praxisgerechte
Verarbeitungsverfahren. Seminars on plastics processing, Institut
für Kunststoffverarbeitung, RWTH Aachen, June 2001.
2. N.S. Ramesh, The heterogeneous nucleation of microcellular foams
assisted by the survival of microvoids in polymers containing low
glass transition particles. Polymer Engineering and Science, 34
(1994) 22, 1698-1706.

Cellular Polymers, Vol. 21, No. 5, 2002 405

4Michaeli paper 405 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

3. F.A. Shutov, Cellular structure and properties of foamed polymers.

Polymeric Foams, Carl Hanser Verlag, Munich, Vienna, New York,
Barcelona, 1991, p. 18-46.
4. M. Shimbo, The viscoelastic behaviour of microcellular plastic with
varying cell size. Polymer Engineering and Science, 35 (1995) 17,
p. 1387-1393.
5. Anon., MuCell extrusion technology for polystyrene. URL:
www.trexel.com, 1998.
6. J.L. Throne and J. Beine, Thermoformen, Werkstoffe - Verfahren
- Anwendungen. Carl Hanser Verlag, Munich, Vienna, 1999.
7. J. Colton, The production of microcellular foams in semicrystalline
thermoplastics. Annual Technical Conference (ANTEC), Society of
Plastics Engineers, Atlanta, 1988, p. 719-722.
8. Anon., Method of providing continuous processing of microcellular
and supermicrocellular foamed materials. US Patent 5866053,
1999.
9. B. Kaufmann, Einfluss des Druckverlaufs im Düsenbereich auf die
Qualität CO 2 -getriebener Schaumfolien. Institut für
Kunststoffverarbeitung, RWTH Aachen, unpublished diploma thesis,
2000, Supervisor: R. Heinz.
10. C.B. Park, A.H. Behravesh and R.D. Venter, Low-density
microcellular foam processing in extrusion using CO2. Polymer
Engineering and Science, 36 (1998) 11, 1812-1823.
11. R.D. Leaversuch, TPO shifts into high gear in automotive instrument
panels. Modern Plastics, 2000, No. 7.
12. C. Hauck, Durchbruch bei Kunststoff-Karosserien - Kunststoffe im
Fahrzeugbau. Plastverarbeiter, 2000, Vol. 51, No. 10.
13. G. Wissing, Direktkaschieren beim Thermoformen - Thermoformen
technischer Teile. SKZ, Würzburg, March 2001.
14. H.W. Wiederoder, Strukturierung der Oberfläche. Thermoformen
technischer Teile. SKZ, Würzburg, March 2001.
15. H.A. Lochner, Umformen von Schaumfolien. Thermoplastische
Schäume - Innovative und praxisgerechte Verarbeitungsverfahren.
Seminars on plastics processing, Institut für Kunststoffverarbeitung,
RWTH Aachen, June 2001.
16. J. Wirtz, Thermoformen von Mehrschichtverbunden,
Thermoplastische Schäume - Innovative und praxisgerechte

406 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 406 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

Verarbeitungsverfahren. Seminars on plastics processing, Institut

für Kunststoffverarbeitung, RWTH Aachen, June 2001.
17. J. Düring and U. Kosok, Hinterspritzen von Folienhauben.
Kunststoffe, Vol. 85 (1995).
18. S. van der Steen, Die Niederdruck-Spritzverfahren. Thermoplastische
Schäume - Innovative und praxisgerechte Verarbeitungsverfahren.
Seminars on plastics processing, Institut für Kunststoffverarbeitung,
RWTH Aachen, June 2001.
19. Anon., Verbesserter Komfort - Emissionsreduzierte und geruchsarme
Kunststoffe. Plastverarbeiter, 2000, Vol. 51, No. 10.
20. R. Peters, Einsatz digitaler Bildverarbeitung zur
Schaumstrukturerkennung. Thermoplastische Schäume - Innovative
und praxisgerechte Verarbeitungsverfahren. Seminars on plastics
processing, Institut für Kunststoffverarbeitung, RWTH Aachen,
June 2001.
21. J. Wirtz, HEIZEN für Windows®, überarbeitetes Benutzerhandbuch
zu HEIZEN für Windows® 5.0, Institut für Kunststoffverarbeitung,
RWTH Aachen, 2001.
22. G. Wissing, Direktkaschieren beim Thermoformen - Fahrzeug-
Innenverkleidungen wirtschaftlich herstellen. Kunststoffe, 1996,
Vol. 86, No. 3.
23. J. Wirtz, Dekorativ und funktionell in einem Schritt - Thermoformen
und Kaschieren. Plastverarbeiter, 2001, Vol. 52, No. 6.
24. G. Fröhlich, Warmformen mit hohem Druck. Kunststoffe, 1997,
Vol. 87, No. 4.
25. Anon., Richtlinie 2000/53/EC des Europäischen Parlamentes und
des Rates vom18th September 2000 über Altfahrzeuge, Bulletin
EU Umwelt [Environment], No. 5, 2000.
26. H. Ehbing, Ansätze zur Charakterisierung der Schaumentstehung,
Rheinish-Westfälische Technische Hochschule Aachen, Thesis,
1999.
27. D. Klempner and K.C. Frisch: Handbook of polymeric foams and
foam technology. Carl Hanser Verlag, Munich 1991.
28. W. Michaeli, Einführung in die Kunststoffverarbeitung. Carl Hanser
Verlag, Munich, 1992.
29. S. Bölinger: Untersuchung verschiedener Einfluβgröβen bei der
Schaumfolienextrusion von Polystyrol mit CO2 als Treibmittel.

Cellular Polymers, Vol. 21, No. 5, 2002 407

4Michaeli paper 407 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

Institut für Kunststoffverarbeitung, RWTH Aachen, unpublished

diploma thesis, 1997, Supervisor: D. Kropp
30. H. Ehbing, Private communication. Bayer AG, Leverkusen, June
1999.
31. C. Hepburn and N. Alam, Influence of fillers on rubber foam
properties. Cell Polymer, 10 (2), p. 99-116, 1991.
32. P. Chaffanjon and G. Verhelst, An automated image analysis
method for the characterisation of flexible foam cellular structure.
Cell Polymer, 11 (1), p. 1-17, 1992.
33. M.H. Rhodes, Image analysis as applied to the characterization of
the cellular structure in urethane foams. Annual Polyurethane
Technical Marketing Conference, Massachusetts, October 1992.
34. N.V. Schwartz and M.T. Bomberg, Image analysis and the
characterization of cellular... In: Plastics Journal of Thermal Insulation,
Vol. 15, October 1991, p. 153-171.
35. A. Hiromasa, H. Teruo and A. Tsuneo, Cell distributions in and
sound absorption characteristics of flexible polyurethane foams.
Journal Appl. Polymer Sci., 65 (7), p. 1395-1402, 1997.
36. F.K. Küderle, Ein Scanner löst Probleme in der Schaumstoffindustrie.
Kunststoffe Synthetics, Yearbook 1993/94, p. 28-31.
37. S.H. Wineland and A.M. Bartz, An instrument for measuring the cell
size of polystyrene and polyethylene foams. Journal of Cellular
Plastics, London, Vol. 22, March 1986, p. 122.
38. B. Lindner, Entwicklung eines Verfahrens zur Analyse von
Schaumstrukturen mittels Bildverarbeitung. Institut für
Kunststoffverarbeitung, RWTH Aachen, unpublished student thesis,
1998, Supervisor: H. Ehbing.
39. T.H. Yu and S.K. Mitra, Efficient approach for the detection of
diffuse edges. Optical Engineering, Vol. 35, No. 12, Dec. 1996, p.
3522-3530.
40. D. Kropp, Private communication, Polymerchemie GmbH, Bad
Sobernheim, July 1999.
41. R. Mohr and B. Triggs, Projective geometry for image analyses. A
tutorial given at ISPRS, Vienna, 1996.
42. W. Michaeli and R. Peters, In-Line Fehlerdetektion bei hohen
Prozessgeschwindigkeiten an extrudierten Folien. Institut für
Kunststoffverarbeitung, RWTH Aachen, Final Report on AiF
Research Project No. 11480N, 2000.

408 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 408 5/11/02, 1:29 pm

 Extruded Foams: Production and Further Processing

SYMBOLS AND ABBREVIATIONS USED

C0 [cm-3] Number of molecules of blowing agent in unit
volume
Chet [cm-3] Number of heterogeneous nucleation sites in
unit volume
f0, f1 [s-1] Frequency of addition of gas molecules to the
nucleus
k [J · K-1] Boltzmann constant
N [cm-3] Cell density
Nf [cm-3] Number of cells per cm3 of foam
Nhet [cm-3s-1] Heterogeneous nucleation rate
Nhom [cm-3s-1] Homogeneous nucleation rate
R [cm] Mean cell diameter
T [K] Temperature, K
Vf [-] Cell volume per cm3
∆G*het [J] Heterogeneous nucleation energy
∆G* [J] Homogeneous nucleation energy

Cellular Polymers, Vol. 21, No. 5, 2002 409

4Michaeli paper 409 5/11/02, 1:29 pm

 W. Michaeli, E. Haberstroh, R. Heinz, J. Wirtz and R. Peters

410 Cellular Polymers, Vol. 21, No. 5, 2002

4Michaeli paper 410 5/11/02, 1:29 pm