Article

Comparison of the Processing of Epoxy Resins in

Pultrusion with Open Bath Impregnation and
Closed-Injection Pultrusion
Sebastian Strauß * , Andreas Senz and Jessica Ellinger
Fraunhofer IGCV, 86159 Augsburg, Germany
* Correspondence: sebastian.strauss@igcv.fraunhofer.de; Tel.: +49-821-90678247




Received: 30 June 2019; Accepted: 13 August 2019; Published: 23 August 2019


Abstract: In this study, the influence of the open bath and closed-injection pultrusion (CIP) processing
methods of epoxy resins on the quality of glass fiber composites was investigated. In addition to
the state-of-the-art epoxy resin system with long pot life, new resin systems with short pot life have
recently been developed. These systems require processing by closed-injection pultrusion. The
epoxies with long pot life allow both processing variants. The experimental work was carried out
with two types of injection and impregnation chambers (ii_chamber), namely with a conical and
a teardrop design. Fully impregnated composites, which were used for further analyses, could
be produced by using the conical ii_chamber. The composite properties of the open bath and the
conical ii_chamber impregnation methods were compared. No significant influence on the bending
stress could be determined; the interlaminar shear strength was up to 10% better with open bath
impregnation than with ii_chamber. For the composites investigated, it was shown that the open
bath and ii_chamber impregnation methods can be used to produce parts with partially comparable
properties, as demonstrated for the epoxy formulation with long pot life. These results indicate
that processing of epoxy systems with a short pot life is also possible by closed-injection pultrusion.
Furthermore, the influence on the composite properties of the time interval between the mixing of an
epoxy resin and processing in an open bath was investigated. No significant effect on the bending
stress and interlaminar shear strength could be determined.

Keywords: pultrusion; epoxy; glass fibers; thermoset; closed-injection pultrusion; open

bath impregnation

1. Introduction
The pultrusion process is an established manufacturing method for obtaining high-quality
fiber-reinforced plastic (FRP) profiles with constant high mechanical properties and offers additional
advantages such as high productivity and comparatively low investment costs. Pultruded composite
profiles are composed of reinforcing fibers (e.g., glass, carbon or basalt) enclosed in a polymer matrix
(e.g., unsaturated polyester, vinyl ester or epoxy). The pultrusion process is a high value-added
process leading to high performance products at low production and energy costs. All of these aspects
have led to an increasing use in a broad variation of applications [1]. The fields of application of
pultrudates range from medical and dental [2–4], infrastructure and construction [5,6] to the transport
and automotive sectors [7]. In addition to the lightweight construction potential of pultrusion profiles,
the general advantages of FRP such as high corrosion resistance and good electrical and thermal
insulation properties are the decisive reasons for their use [1,5]. The properties of pultrudates depend
not only on the raw materials (fibers and matrix) composition but also strongly on the processing
history [8,9].

J. Compos. Sci. 2019, 3, 87; doi:10.3390/jcs3030087 www.mdpi.com/journal/jcs

J. Compos. Sci. 2019, 3, 87 2 of 11

The use of epoxy resin systems is increasing because of the good achievable properties of
the composites e.g., high glass transition temperature (Tg) and good flex fatigue properties [10].
Anhydride-curing epoxy resin systems with a long pot life are usually processed in pultrusion with
open impregnation baths, but they could also be processed with injection and impregnation chambers
(ii_chamber). The state of the art is the open processing of standard resins in pultrusion due to their
long pot life [11]. Epoxy resins are, therefore, also processed mainly in the open process. Closed
impregnation processes are an alternative for the reduction of the environmental impact on the resin
bath, the reduction of emissions into the environment and the reduction of open resin components and
impregnated fiber strands [12]. In addition, by employing closed-injection pultrusion (CIP), highly
reactive amine curing epoxy systems with a short pot life of only a few minutes can also be used. The
processing of these highly reactive resin systems is not yet widespread.
One of the most important parameters in pultrusion is the viscosity of the resin [13–17]. Other
influencing factors that are frequently investigated in further studies—besides the pulling speed—are
the fiber volume fraction (FVF), curing temperature, reaction kinetics with resin shrinkage and the
compression ratio at the injection point [16,18]. In many of these studies, only theoretical investigations
are undertaken [19]. No comparative studies of an epoxy resin with CIP and open bath impregnation
are known.
This study investigated how the mechanical properties of the composites behave due to the
different impregnation processes. This was studied experimentally in the context of the present
work. For this purpose, pultrusion trails were run with two different ii_chamber inner geometries.
Furthermore, the influence of the time span between the mixing and use of the resin system in the open
bath impregnation system on the mechanical performance of the composite profiles was investigated.
It was assumed that a slowly increasing viscosity of the resin system induced by a longer holding
time influences the wetting of the fibers. Furthermore, the influence of different pulling-speeds on the
mechanical characteristics of the pultrudates was investigated.

2. Materials and Methods

The term pultrusion refers to the process principle: reinforcing fibers impregnated with a matrix
resin are pulled through a heated die. In the die, the thermosetting matrix material cures through
polymerization and the composite is created in the desired shape in one step. The process for the
production of composite profiles with a constant cross-section was developed by W.B. Goldsworthy in
the United States in the 1950s [20]. For the production of high-performance composite profiles using
the pultrusion process, it is necessary to achieve a uniform impregnation of uniaxial and multiaxial
textiles and semi-finished products with the matrix material. There are various known open and
closed process variants for this purpose [21]. For every system, there are some challenges and also
some disadvantages in their use [22,23]. Figure 1 schematically shows the pultrusion process with
open bath impregnation. The principal steps of the process are represented by: (1) roving and textile
provision with preforming, (2) impregnation (process variant dip bath), (3) die for curing, (4) take-off
unit and packaging.
Figure 2 shows the basic impregnation mechanisms for pultrusion applications schematically.
The two open bath characteristics are shown in the upper part of the diagram: (a) describes the
straight-through process and (b) the dip bath. In (c)–(f) the process variants of the CIP are shown:
teardrop (c), high-pressure (d), conical (e) and (f) siphon.
The number of factors influencing composite quality in pultrusion technology is significant [1,8].
The intensity of influence varies [24]. Based on the main influencing variables mentioned above, this
study will investigate the temporal influence of the viscosity development for epoxy and its influence
on the composite quality. A comparison between two impregnation methods for epoxy will also
be investigated: open bath and CIP. The influencing parameters FVF and curing temperature were
developed and optimized in preliminary studies and should remain constant. The reaction kinetics
and material shrinkage parameters are determined by the material selection. In the preliminary work,
 production of composite profiles with a constant cross-section was developed by W.B. Goldsworthy
in the United States in the 1950s [20]. For the production of high-performance composite profiles
using the pultrusion process, it is necessary to achieve a uniform impregnation of uniaxial and
multiaxial textiles and semi-finished products with the matrix material. There are various known
open and
J. Compos. closed
Sci. 2019, 3, 87process variants for this purpose [21]. For every system, there are some challenges 3 of 11
and also some disadvantages in their use [22,23]. Figure 1 schematically shows the pultrusion process
with open bath impregnation. The principal steps of the process are represented by: (1) roving and
ii_chambers were developed
textile provision that have
with preforming, good processing
(2) impregnation properties.
(process variantThe
dip influence of the
bath), (3) die for compression
curing, (4)
ratio is, therefore,
take-off unit andnot investigated in this study.
packaging.

J. Compos. Sci. 2019, 3, x 3 of 11

Figure 2 shows the basic impregnation mechanisms for pultrusion applications schematically.
The two open bath characteristics are shown in the upper part of the diagram: (a) describes the
straight-through process and (b) the dip bath. In (c)–(f) the process variants of the CIP are shown:
teardrop (c), high-pressure (d),Figure
conical
Figure1.1.(e) and
Basic
Basic (f) siphon.
elements
elements of aa pultrusion
pultrusionline.
line.

Figure
Figure 2.
2. Impregnation
Impregnation mechanisms
mechanisms for
for pultrusion.
pultrusion.

The number
The experiments wereinfluencing
of factors carried outcomposite
in two sections.
qualityIn in the first experimental
pultrusion technology step, the fiber[1,8].
is significant was
impregnated
The intensity of with an openvaries
influence bath;[24].
in the experimental
Based on the mainstepinfluencing
two, ii_chambers
variables were used to impregnate
mentioned above, this
the fibers. For all tests, a 120 × 4 mm 2 flat profile pultrusion die was used. The fiber package of the
study will investigate the temporal influence of the viscosity development for epoxy and its influence
composite
on was defined
the composite quality.as A
a unidirectional
comparison betweenglass-fiber
tworeinforcement. For all experiments,
impregnation methods for epoxy will184 also
PS4100
be
single-end rovings
investigated: from and
open bath Owens CIP.Corning (Toledo, OH,
The influencing USA) with
parameters FVFa fineness
and curing of 4800 tex were were
temperature used.
This resultsand
developed in fiber volumeincontent
optimized for thestudies
preliminary composite
andofshould
70%. One anhydride-curing
remain constant. The pultrusion epoxy
reaction kinetics
resinmaterial
and system shrinkage
from each parameters
Huntsman (The Woodlands,
are determined byTX,
theUSA) andselection.
material Sika (Baar, InSwitzerland)
the preliminary waswork,
used
(see Table 1).were
ii_chambers All ofdeveloped
the tests were
that carried
have goodout processing
at two pull-off speeds (0.3
properties. Theand 0.5 m/min);
influence of thethe pultrusion
compression
die was
ratio tempered not
is, therefore, with four heating
investigated inzones (see Table 2).
this study.
The experiments were carried out in two sections. In the first experimental step, the fiber was
Table 1. Resin systems.
impregnated with an open bath; in the experimental step two, ii_chambers were used to impregnate
the fibers. For
Material all tests, a 120 × 4
Componentmm² flat profile
Manufacturer die was Name
pultrusion used. The fiber packageParts of the
composite was defined as a unidirectional glass-fiber reinforcement. For all experiments, 184 PS4100
resin Huntsman XB 3585 100
single-end rovings from Owens Corning (Toledo,
hardener OH, USA) withAradur
Huntsman a fineness
917CH of 4800 tex were
80 used.
EP 1
This results in fiber volume content for the composite
accelerator Huntsman of 70%.DY-080/ACC-960
One anhydride-curing 2.5 pultrusion
epoxy resin system from eachIMR Huntsman (TheChem-Trend
Woodlands, TX, Chemlease
USA) andIC 25 (Baar, Switzerland)
Sika 6
was used (see Table 1). All ofresin the tests were carried SIKAout at two pull-off
Biresin speeds
CR141 (0.3 and 0.5 100 m/min);
the pultrusion die was hardener
tempered with four heatingSIKA
zones (see TableBiresin
2). CH141 90
EP 2
accelerator SIKA Biresin CA141 2
IMR Table 1.Chem-Trend
Resin systems. Chemlease IC 25 6

Material Component Manufacturer Name Parts

resin Huntsman XB 3585 100
hardener Huntsman Aradur 917CH 80
EP 1
accelerator Huntsman DY-080/ACC-960 2.5
J. Compos. Sci. 2019, 3, 87 4 of 11

Table 2. Design of experiments: used resin, mixing time and impregnation method.

Time after Heating Die Speed

No. Impregnation Resin
Mixing (h) (Z1/Z2/Z3/Z4) (m/min)
1A dip bath EP 1 0 160/180/170/170 0.3
1A dip bath EP 1 0 160/180/170/170 0.5
1B dip bath EP 1 3 160/180/170/170 0.3
1B dip bath EP 1 3 160/180/170/170 0.5
1C dip bath EP 1 5 160/180/170/170 0.3
1C dip bath EP 1 5 160/180/170/170 0.5
1D dip bath EP 2 0 180/180/180/180 0.3
1D dip bath EP 2 0 180/180/180/180 0.5
1E dip bath EP 2 3 180/180/180/180 0.3
1E dip bath EP 2 3 180/180/180/180 0.5
1F dip bath EP 2 5 180/180/180/180 0.3
1F dip bath EP 2 5 180/180/180/180 0.5
2G conical EP 1 0 160/180/170/170 0.3
2G conical EP 1 0 160/180/170/170 0.5
2H * teardrop EP 1 0 160/180/170/170 0.3
2H * teardrop EP 1 0 160/180/170/170 0.5
2I conical EP 2 0 180/180/180/180 0.3
2I conical EP 2 0 180/180/180/180 0.5
2J * teardrop EP 2 0 180/180/180/180 0.3
2J * teardrop EP 2 0 180/180/180/180 0.5
* The 2J trials were planned with the same test assembly but could not be successfully carried out due to processing
instability. As a consequence, the 2H experiments were canceled. See Section 3.4.

In the first trial section, an open bath in the dipping bath variant was used to impregnate the fiber
package (see diagram in Figure 2b). In order to investigate the influence on the impregnation quality
of the time span after mixing of the resin system, the resin was used in three time gradations after
mixing: immediately after mixing (0 h), three hours after mixing and five hours after mixing. Table 2
gives an overview of the design of the experiments with the names of the test series.
To impregnate the fibers in the second part of the experiment, ii_chambers were used that
were directly attached to the die. The fibers were impregnated in an ii_chamber with conical and
teardrop geometry (1) and then pulled through a heated die of 1000 mm length and 4 tempering zones
(2) (see Figure 3). A composite profile was continuously pulled out of the die (3). The resin was injected
by means of a pressure pot with three injection nozzles located at the top and bottom sides of the
ii_chamber (4). 3, x
J. Compos. Sci. 2019, 5 of 11

Figure
Figure 3. Pultrusion setup.
3. Pultrusion setup.

For the study, one pultrusion trial was performed per experimental setup. When the pulling
3. Results and Discussion
speed was changed, Pultrusion was performed for 20 min to achieve static process conditions again.
The monitored
This was results of by
themeans
mechanical testing heating
of the required of the power
composite areoccurred
and the summarized below.
pull force. These are
Subsequently,
three meters of profile were taken for materials testing. The test specimens were taken from theascenter
analyzed on the basis of interlaminar shear strength and bending strength of the composite well
as each
of by optical microscopic
composite profile. analysis.
Five valid specimens were tested for each mechanical test.

3.1. Pultrusion with Open Bath Impegnation

When considering interlaminar shear strength, the two resin systems exhibit different behavior.
The ILSS values of the EP 1 resin remain constant over both the variation of the speed and the
different processing times after the mixing of the resin. There is no significant influence of the time
J. Compos. Sci. 2019, 3, x 5 of 11

J. Compos. Sci. 2019, 3, 87 5 of 11

The laminate quality of the pultrudates was evaluated on the basis of mechanical properties.
These were determined by the three-point bending test according to DIN EN ISO 14125 and the
interlaminar shear strength (ILSS) test according to DIN EN ISO 14130. In accordance with the norms,
the dimensions of the test specimens were adapted to a thickness of 4 mm. Rheological measurements
Figure 3. Pultrusion setup.
were carried out to determine the influence of viscosity on the longer pot life of the resin systems.

3. Results
3. Results and
and Discussion
Discussion
The results
The resultsofofthethe mechanical
mechanical testing
testing ofcomposite
of the the composite are summarized
are summarized below.
below. These areThese are
analyzed
analyzed
on onofthe
the basis basis of interlaminar
interlaminar shear
shear strength andstrength
bendingand bending
strength strength
of the of theascomposite
composite well as by as well
optical
as by optical microscopic
microscopic analysis. analysis.

3.1. Pultrusion with Open Bath Impegnation

When considering interlaminar shear strength, strength, the
the two
two resin
resin systems
systems exhibit
exhibitdifferent
different behavior.
behavior.
The ILSS
The ILSSvalues
valuesofofthethe EPEP 1 resin
1 resin remain
remain constant
constant over the
over both both the variation
variation of theand
of the speed speed and the
the different
different processing
processing times after times after theofmixing
the mixing of the
the resin. resin.
There There
is no is no significant
significant influenceinfluence
of the timeof between
the time
betweenofmixing
mixing the resinof system
the resin system
and and the processing
the processing point. Variations
point. Variations of the characteristic
of the characteristic values arevalues
in the
are in of
range thethe
range of thedeviation.
standard standard deviation.
When considering the characteristics
characteristics of of the EP 2 resin with regard to the pull-off speed and
processing time of the resin, two effects
processing effects are
are seen.
seen. The characteristic values for the composite with the
pull-off speed
higher pull-off speedof of0.5
0.5m/min
m/minshow showhigher
highercharacteristic
characteristicvalues
values over
over allall three
three processing
processing times
times of
of the resin. This is most evident when the resin is processed immediately
the resin. This is most evident when the resin is processed immediately after mixing. Here, the ILSSafter mixing. Here, the
ILSS parameter
parameter is moreisthanmore 10%than 10%than
higher higher than the reached
the parameter parameter reached
at 0.3 m/min.atThe 0.3effect
m/min. The effect
decreases over
decreases
the overtime
increasing the increasing
between mixingtime between mixing of
and processing and
theprocessing of the 4).
resin (see Figure resinIn (see Figure
addition, 4). In
a slight
addition,ina ILSS
increase slightproperties
increase in ILSS
over theproperties
processingovertimethe processing
after mixing oftimethe after
resin mixing of the resin EP 2
EP 2 is seen.
is seen.

Interlaminar shear
Figure 4. Interlaminar shear strength for different pulling speeds and resin mixing times.
times.

When
When considering
consideringbending
bendingstiffness
stiffnessinin
terms of the
terms pulling
of the speed
pulling and and
speed processing time time
processing of theofresin
the
after
resin mixing, no trend
after mixing, no of the characteristic
trend values can
of the characteristic be found;
values can be the values
found; thecan be regarded
values as constant
can be regarded as
for both resin systems. Variations are in the range of the standard deviation (see Figure 5).
constant for both resin systems. Variations are in the range of the standard deviation (see Figure 5).
J. Compos. Sci. 2019, 3, 87 6 of 11
J.J. Compos.
Compos. Sci.
Sci. 2019,
2019, 3,
3, xx 66 of
of 11
11

Figure
Figure 5. Bending
Bending stress
stress for
for different
different pulling speeds and resin mixing
mixing times.
times.

A long-term measurement over seven hours at room temperature using a plate-plate rheometer
was carried
was carried out out to
to check
checkthe theviscosity
viscositycurve
curveafter
afterformulation
formulationofofthe theresin
resinover
over time.
time. TheThe viscosity
viscosity at
at the
the beginning
beginning of the
of the measurements
measurements waswas about
about 3200
3200 mPa·s
mPa·s for for
resinresin
EP EP 1 and
1 and 14001400 mPa·s
mPa·s for for
resinresin
EP
2EP 2 (see
(see Figure
Figure 6). A6). A constant
constant increase
increase in viscosity
in viscosity can becan befor
seen seenEPfor EP 1.
1. The The viscosity
viscosity of resinofEPresin EP 1
1 almost
almost doubled
doubled to 7358to 7358 after
mPa·s mPa·sthree
afterhours.
three hours.
After aAfter
timeaoftimefiveofhours,
five hours, a viscosity
a viscosity of 10,650
of 10,650 mPa·s mPa·s
was
was measured.
measured. A lowerA lower
initialinitial viscosity
viscosity was measured
was measured for theforEPthe EP 2 resin.
2 resin. The increase
The increase overistime
over time lower is
lower than for EP 1. For EP 2, a doubling to 2700 mPa·s was measured after 3 h
than for EP 1. For EP 2, a doubling to 2700 mPa·s was measured after 3 h from the initial value. Afterfrom the initial value.
After five hours,
five hours, the viscosity
the viscosity was was
28722872 mPa·s.
mPa·s. TheThe increase
increase in viscosity
in viscosity of of theEP
the EP2 2resin
resinisis noticeably
noticeably
slower. The characteristic of the complex viscosity is more complex for for EP 2 than for EP 1, which can
be approximated linearly.
linearly.

Figure 6. Complex
Complex viscosity
viscosity long-time
long-time measurement at room temperature.
J. Compos. Sci. 2019, 3, 87 7 of 11

J. Compos. Sci. 2019, 3, x 7 of 11

3.2. Closed-Injection Pultrusion (CIP)
3.2. Closed-Injection Pultrusion (CIP)
The results of the material analysis of the composites produced with CIP are analyzed as follows.
The results of the material analysis of the composites produced with CIP are analyzed as follows.
3.2.1. CIP withSci.ii_Chamber
J. Compos. 2019, 3, x with Inner Teardrop Geometry 7 of 11
3.2.1. CIP with ii_Chamber with Inner Teardrop Geometry
The3.2.test series for Pultrusion
Closed-Injection CIP with(CIP)ii_chamber with inner teardrop geometry was started with EP 2.
It wasThe
started
test with
seriesa for
pull-off speedii_chamber
CIP with of 0.3 m/min. withDuring optical testing,
inner teardrop geometryvarious
was defects
started were
with detected
EP 2. It
The results of the material analysis of the composites produced with CIP are analyzed as follows.
macroscopically
was started withacross a pull-offthe speed
profileofcross-section.
0.3 m/min. DuringNo uniform fiber impregnation
optical testing, various defects waswere
achieved.
detectedAt
amacroscopically
pull-off speed
3.2.1. CIP withacross
of 0.5 the profile
m/min,
ii_Chamber cross-section.
the Inner
with fibers were only
Teardrop No impregnated
uniform fiber at
Geometry impregnation
the boundary, waswhile
achieved. At a
the center
ofpull-off speedwas
the profile of 0.5
notm/min, the fibers
impregnated withwere only
resin impregnated
(see Figure 7). Due at the
to boundary, whileimpregnation,
the incomplete the center of
The test series for CIP with ii_chamber with inner teardrop geometry was started with EP 2. It
theprofiles
the profile was
werenot
was started notimpregnated
with analyzed with
further
a pull-off speed resin (seeDuring
bym/min.
of 0.3 means Figure 7). Due
of material
optical to the
testing.
testing, incomplete
These
various results
defects impregnation,
match
were the
the results
detected
profiles
from were not analyzed
preliminary
macroscopically experiments further
across the andby
profile themeans of material
description
cross-section. Nofrom testing.
uniform thefiber These
patent byresults match
describing
impregnation was theuse
the
achieved.results from
ofatextiles
At
preliminary
when pull-offexperiments
teardrop ii_chambers
speed and
arethe
of 0.5 m/min, thedescription
used [25]. were
fibers fromimpregnated
The only
second the patent
series bythedescribing
of pultrusion
at teststhe
boundary, usethe
with
while of center
EP 1textiles
(2H)ofwaswhen
not
teardropthe
carried out. ii_chambers
profile was not are used
impregnated[25]. The
with second
resin (see series
Figure of
7). pultrusion
Due to the tests
incompletewith EP 1 (2H)
impregnation, was
the not
carriedprofiles
out. were not analyzed further by means of material testing. These results match the results from
preliminary experiments and the description from the patent by describing the use of textiles when
teardrop ii_chambers are used [25]. The second series of pultrusion tests with EP 1 (2H) was not
carried out.

Figure7.7. Composites
Figure Compositesmade
madewith
withclosed-injection pultrusion
closed-injection (CIP)
pultrusion using
(CIP) teardrop
using ii_chamber
teardrop at 0.3at
ii_chamber
m/min (two top profiles) and 0.5 m/min. dry spots are marked with arrows.
0.3 m/min (two top profiles) and 0.5 m/min. dry spots are marked with arrows.
Figure 7. Composites made with closed-injection pultrusion (CIP) using teardrop ii_chamber at 0.3
3.2.2. m/min (two top profiles) and 0.5 m/min. dry spots are marked with arrows.
3.2.2.CIP
CIPwith
with ii_Chamber
ii_Chamber withwith Inner Conical Geometry
The3.2.2.
The CIP with ii_Chamber
characteristics
characteristics of
of the with Inner Conical
the interlaminar
interlaminar shearGeometry
strength of
strength of the
the resin
resin when
whenprocessed
processedwith withCIP
CIPand
and
conical ii_chamber
conical ii_chamber are similar
are similar
The characteristics for both resin
forinterlaminar
of the systems. The
both resin shear strength characteristic
The of
characteristic
the resin whenvalues
values decrease
decrease
processed with CIP slightly
slightly with
and with
increasing
increasing pull
conicalpullspeed.
speed. For
ii_chamber are both
For both
similar resin systems,
for both a decrease
resin systems. of
The of about 4%
about 4%can
characteristic canbebedecrease
values measured
measured inincomparison
slightly comparison
with
totothe increasing
thelower pullpull
lowerpull speed.
speed,
speed, seeFor
see both resin
Figure
Figure 8. systems, a decrease of about 4% can be measured in comparison
8.
to the lower pull speed, see Figure 8.

Figure 8. Interlaminar shear for different pulling speeds and resin mixing times.
Figure 8. Interlaminar shear for different pulling speeds and resin mixing times.

When considering bending stiffness, the increase in pulling speed cannot be used to identify
When considering
Figure bending shear
8. Interlaminar stiffness,
for the increase
different in pulling
pulling speedsspeed
andcannot be usedtimes.
resin mixing to identify any
any changes
changesininthe
thecharacteristic valuesthat
characteristic values thatareare not
not within
within the the range
range of standard
of the the standard deviation
deviation (see (see
FigureWhen
9).
Figureconsidering
9). bending stiffness, the increase in pulling speed cannot be used to identify any
changes in the characteristic values that are not within the range of the standard deviation (see
Figure 9).
J. Compos. Sci. 2019, 3, 87 8 of 11
J. Compos. Sci. 2019, 3, x 8 of 11

J. Compos. Sci. 2019, 3, x 8 of 11

Figure 9. Bending
Figure stress
9. Bending stressfor
fordifferent
different pulling speedsand
pulling speeds and resin
resin mixing
mixing times.
times.

3.3. Microscopy
3.3. Microscopy
For aFormore in-depth
a more in-depthexamination
examinationofofthe thecomposite, microsections
composite, microsections areare made
made andand examined
examined with with
an incident light microscope. An exemplary micrograph of the pultrudates
an incident light microscope. An exemplary micrograph of the pultrudates with conical CIP and with conical CIP and open
openbath
bathproduction
production is shown
is shown for each examination.
for each Figure Figure
examination. 10 shows 10ashows
pictureaofpicture
the experiment 1A-03
of the experiment
on the left. The composite was produced by open bath impregnation. On the right, a picture of the
1A-03 on the left. The composite was produced by open bath impregnation. On the right, a picture
test series 2G-03 with the processing method CIP with conical ii_chamber can be seen. In the
of the test series 2G-03 with the processing method CIP with conical ii_chamber can be seen. In the
comparison of the images, it is noticeable that significantly more pores—marked with arrows as an
comparison Figure
of the 9. Bending
images, stress for different
it is noticeable pulling speeds
that significantly more andpores—marked
resin mixing times. with arrows as an
example—are visible in the laminate in the CIP composite, the boundary surfaces of the individual
example—are visible in the laminate
roving bundles—marked in the
with brackets in CIP composite,
the left picture—arethe noboundary surfaces
longer visible. of the
These individual
are clearly
3.3. Microscopy
roving bundles—marked
visible in the left image,with
whereasbrackets in theofleft
the number picture—are
voids is much lower. no longer visible. These are clearly
visible
Forinathe leftin-depth
more image, whereas the number
examination of voids is much
of the composite, lower. are made and examined with
microsections
One explanation
an incident for the An
light microscope. no exemplary
longer visible bundles of the
micrograph roving in the CIP
pultrudates withcould be the
conical CIPincreased
and open
pressure ratios in is
bath production the ii_chamber.
shown for each The filaments ofFigure
examination. the rovings coulda have
10 shows been
picture of pressed more strongly
the experiment 1A-03
into
on the left. The composite was produced by open bath impregnation. On the right, a picture of the
each other by this. The boundaries of the roving bundles were overcome; the distribution
filaments
test seriesin2G-03
the composite
with theappearsprocessinguniform.
methodThereCIPis air
with between
conicaland in the drycan
ii_chamber rovings of theInfiber
be seen. the
package. This air is drawn into the ii_chamber. The impregnation of the
comparison of the images, it is noticeable that significantly more pores—marked with arrows as an fibers takes place under
pressure over avisible
example—are short distance in the ii_chamber.
in the laminate in the CIPDue to the flow
composite, theresistances
boundaryof the resin
surfaces ofand
the the pull-off
individual
movement of the fibers andwith
roving bundles—marked the resin,
bracketsthe in
airthe
could
leftnot completelyno
picture—are escape
longer to visible.
the inletThese
side. This could
are clearly
have made trapped air bubbles visible as pores in the
visible in the left image, whereas the number of voids is much lower. composite.

Figure 10. Microsections of open bath (left) und conical ii_chamber (right) composites.

One explanation for the no longer visible bundles of roving in the CIP could be the increased
pressure ratios in the ii_chamber. The filaments of the rovings could have been pressed more strongly
into each other by this. The boundaries of the roving bundles were overcome; the distribution of the
filaments in the composite appears uniform. There is air between and in the dry rovings of the fiber
package. This air is drawn into the ii_chamber. The impregnation of the fibers takes place under
pressure over a short distance in the ii_chamber. Due to the flow resistances of the resin and the pull-
off movement of the fibers and the resin, the air could not completely escape to the inlet side. This
could have made trapped air bubbles visible as pores in the composite.

Figure 10. Microsections of open bath (left) und conical ii_chamber (right) composites.
composites.

3.4. Discussion
One explanation for the no longer visible bundles of roving in the CIP could be the increased
pressure ratios
In the first in theofii_chamber.
part the study, itThe
wasfilaments of thehow
investigated rovings could havebehavior
the processing been pressed more
of epoxy strongly
resin over
into each other by this. The boundaries of the roving bundles were overcome; the distribution
time affects the properties of pultruded composites after mixing. Two anhydride-curing resins were of the
filaments in the composite appears uniform. There is air between and in the dry rovings of the fiber
package. This air is drawn into the ii_chamber. The impregnation of the fibers takes place under
pressure over a short distance in the ii_chamber. Due to the flow resistances of the resin and the pull-
off movement of the fibers and the resin, the air could not completely escape to the inlet side. This
could have made trapped air bubbles visible as pores in the composite.
 J. Compos. Sci. 2019, 3, x 9 of 11

J.3.4. Discussion
Compos. Sci. 2019, 3, 87 9 of 11

In the first part of the study, it was investigated how the processing behavior of epoxy resin over
time
used for affects
thisthe properties
purpose. Theof pultrudedofcomposites
evaluation after mixing.
the mechanical Twoshowed
properties anhydride-curing
that the time resins were
influence
used for this purpose. The evaluation of the mechanical properties showed
between 0–5 h is only partially visible in the characteristic values. No influence on the bending stressthat the time influence
between
could 0–5 h is only When
be determined. partially visible in
analyzing the
the characteristic
influence values. No shear
on interlaminar influence on theanbending
strength, stress
improvement
could be determined. When analyzing the influence on interlaminar shear
of the properties over the longer period between mixing and processing was observed for one of the strength, an improvement
of the
two properties
resins over the longer period between mixing and processing was observed for one of the
investigated.
two resins investigated.
It was shown by rheological measurement that viscosity increases continuously with time after
It was shown by rheological measurement that viscosity increases continuously with time after
mixing of the resin. This is confirmed by the observation of resin behavior in the open impregnation
mixing of the resin. This is confirmed by the observation of resin behavior in the open impregnation
bath from previous experiments. The assumption that this could make the impregnation of the filaments
bath from previous experiments. The assumption that this could make the impregnation of the
more difficult was not confirmed for the materials investigated on the basis of the characteristic values
filaments more difficult was not confirmed for the materials investigated on the basis of the
used. On the contrary, for resin EP 2, slightly increasing ILSS parameters were determined with
characteristic values used. On the contrary, for resin EP 2, slightly increasing ILSS parameters were
increasing time after mixing. No influence on the flexural strength of the two variables investigated,
determined with increasing time after mixing. No influence on the flexural strength of the two
mixture processing time and pulling speed, could be found.
variables investigated, mixture processing time and pulling speed, could be found.
The hypothesis that the quality of the composite deteriorates for a time interval between mixing
The hypothesis that the quality of the composite deteriorates for a time interval between mixing
and processing could not be established for the investigated parameters. No clear trend of the time
and processing could not be established for the investigated parameters. No clear trend of the time
interval affecting the laminate quality could be found for the two resins that were investigated. It was
interval affecting the laminate quality could be found for the two resins that were investigated. It was
found that for both resin systems for the period investigated up to five hours, that the characteristic
found that for both resin systems for the period investigated up to five hours, that the characteristic
values—which were generated immediately after the mixing of the resin—can be used as a reference
values—which were generated immediately after the mixing of the resin—can be used as a reference
for
foraamaterial
materialororcomponent
component design. For both
design. For both systems,
systems, nonodeterioration
deteriorationof ofthe
theinvestigated
investigatedparameters
parameters
was
wasfoundfoundwithin
withinthe theinvestigated
investigated time
time period.
period.
InIn order to work on the hypothesis of
order to work on the hypothesis the influence
of the influence ofof the
theimpregnation
impregnationmethod,method,the thecomparison
comparison
ofofthethecomposite
composite properties was investigated
properties by different
was investigated by impregnation approaches.approaches.
different impregnation The experimentalThe
work was carried out for the two aforementioned epoxy systems with
experimental work was carried out for the two aforementioned epoxy systems with three different three different impregnation
mechanisms.
impregnationAmechanisms.
direct comparison of thecomparison
A direct properties of of the
theopen bath and
properties ofCIPtheimpregnation
open bath and methods
CIP
with conical andmethods
impregnation teardropwithii_chamber
conical enabled the drawing
and teardrop of two enabled
ii-chamber conclusions. the It was shown
drawing that
of two
the
conclusions. It was shown that the ii_chamber impregnation method with teardrop design resultedof
ii_chamber impregnation method with teardrop design resulted in insufficient impregnation
unidirectional
in insufficientfiber stacks withoftheunidirectional
impregnation EP 2 resin formulation.
fiber stacks The comparison
with the EP 2ofresinthe properties
formulation. between
The
open bath impregnation
comparison and CIP
of the properties conical
between ii_chamber
open shows an up
bath impregnation andto CIP
8% decrease in the bending
conical ii-chamber showsstress
an
values for the CIP impregnation case. When comparing the ILSS characteristic
up to 8% decrease in the bending stress values for the CIP impregnation case. When comparing the values at the higher
pulling speeds of 0.5
ILSS characteristic m/min,
values property
at the losses ofspeeds
higher pulling more of than
0.5 15%
m/min,became apparent
property lossescompared
of more than to 15%
open
bath
became impregnation.
apparent comparedThis trend to can
openalso
bathbeimpregnation.
observed for the Thisother
trendILSS
can parameters.
also be observedThe characteristic
for the other
values of the CIP are
ILSS parameters. lower
The than those values
characteristic of the open
of thebath
CIP impregnation
are lower than (see Figure
those of11).theThe increased
open bath
porosity
impregnationin the ii_chamber
(see Figurecomposites could be porosity
11). The increased the reason in for
thethe reduced properties
ii_chamber composites compared
could betothethe
open
reason bath
forimpregnation method. compared to the open bath impregnation method.
the reduced properties

Figure11.
Figure 11. Comparison
Comparison of pultrusion
pultrusion tests by bending
bending stress
stress (left)
(left) and
and interlaminar
interlaminarshearshearstrength
strength
(right):Impregnation
(right): Impregnationwith
withopen
open bath
bath (label
(label 1) and
1) and CIPCIP (label
(label 2): Mixing
2): Mixing timetime
of theofresin
the resin directly
directly before
before
the the experiments
experiments with
with two two pulling
pulling speeds speeds
(0.3 and(0.3
0.5and 0.5 m/min)
m/min)
J. Compos. Sci. 2019, 3, 87 10 of 11

4. Conclusions
This study investigated the influence of different impregnation methods on the pultrusion process
for anhydride curing epoxy resins. The time interval between mixing and processing of the resin was
studied for the open bath impregnation case. No effect could be determined from the ILSS results for
the EP 1 resin formulation; for EP 2, the ILSS values increased with processing time and speed. No
effect on the bending stress could be determined for either system by either influencing variable.
Furthermore, this study focused on the research of CIP with two different ii_chambers. It was
observed that the teardrop ii_chamber used in the experiments leads to insufficient impregnation of a
unidirectional roving stack with the studied resin systems. This observation is in agreement with the
descriptions in the patent specification [25]. The processing with a conical ii_chamber enabled a stable
pultrusion process. The comparison of the properties with the open bath experiments shows reduced
properties in bending stress and interlaminar shear strength. The characteristic values obtained—with
impregnation with the conical ii_chamber—of bending stress of 1550 MPa (EP 1; 0 h; 0.3 m/min) and
1481 MPa (EP 2; 0 h; 0.3 m/min) are comparable with the values of 1650 MPa from the literature [26].
The same resin system as EP 2 was used for this, but the glass fibers used differed. No statement is
made about the geometry of the conical ii_chamber in [26].
From the results, it can be deduced that when processing resin systems with long pot lives
with glass fibers in the CIP process, slight property decreases may be expected in comparison to the
open bath impregnation. Furthermore, it was seen how sensitive materials react to process variants
with different impregnation processes. This insight can be used to improve the assembly design and
consequently the composite part performance according to the material system to be processed.

Author Contributions: Conceptualization, S.S.; methodology, S.S.; validation, A.S., J.E.; formal analysis, A.S., J.E.;
investigation, A.S., J.E.; data curation, A.S., J.E.; writing—original draft preparation, S.S., A.S.; writing—review
and editing, S.S.; visualization, J.E.; project administration, S.S.; funding acquisition, S.S.
Funding: This research received no external funding. The publication was funded by the Hanns-Seidel-Foundation
as part of a scholarship.
Conflicts of Interest: The authors declare no conflict of interest.

References and Notes

1. Starr, T.F. Pultrusion for Engineers, 1st ed.; Woodhead Publishing Ltd.: Cambridge, UK, 2000; pp. 1–17.
2. Lange, S.C. Miniaturisierung des Strangziehverfahrens am Beispiel der Herstellung von Kanülen aus
kohlenfaserverstärktem Kunststoffen. Ph.D. Thesis, RWTH Aachen University, Aachen, Germany, 2005.
3. Schmitz, S. Strangziehen von MR-sicheren Führungsdrähten aus Faserverbundkunststoffen. Ph.D. Thesis,
RWTH Aachen University, Aachen, Germany, 2010.
4. Xu, X.; He, L.; Zhu, B.; Li, J.; Li, J. Advances in polymeric materials for dental applications. Polym. Chem.
2017, 8, 807–823. [CrossRef]
5. Röchling high-performance plastic gruppe, composite profiles, 2014.
6. Let’s talk pultrusion. Aliancys AG, Schaffhausen, Switzerland, 2016.
7. Geiger, R.; Pahl, J. CFK-Strukturen in der automobilen Serienfertigung. Lightweight Des. 2017, 10, 52–55.
[CrossRef]
8. Meyer, R.W. Handbook of Pultrusion Technology, 1st ed.; Chapman and Hall Ltd.: Boston, MA, USA, 1986.
9. Strauß, S.; Meisenheimer, R.; Hecking, A.; Wilhelm, S.A. Demonstration of manufacturing predominantly
bio-based composite profiles with pultrusion, using cellulose fibre and aliphatic polyurethane. In Proceedings
of the 22nd International Conference on Composite Materials, Melbourne, Australia, 11–16 August 2019.
10. Watkins, M. Development of new anhydride curing agents for high performance epoxy pultrusion.
In Proceedings of the 14th World Pultrusion Conference, Vienna, Austria, 1–2 March 2018.
11. Engelen, H. Development and comparison of resin injection and impregnation chambers for pultrusion.
In Proceedings of the 11th World Pultrusion Conference, Istanbul, Turkey, 22–23 March 2012.
J. Compos. Sci. 2019, 3, 87 11 of 11

12. Engelen, H. Pultrusion mit duromeren Harzsystemen: Aktuelle Entwicklungen im Bereich der Injektions-
Pultrusion und ausgewählte Anwendungs-Beispiele. In Proceedings of the Composite-Customer-Days SIKA,
Munich, Germany, 11–12 July 2018.
13. Shakya, N.; Roux, J.A.; Jeswani, A.L. Effect of resin viscosity in fiber reinforcement compaction in resin
injection pultrusion process. Appl. Compos. Mater. 2013, 20, 1173–1193. [CrossRef]
14. Masuram, N.B.; Roux, J.A.; Jeswani, A.L. Resin viscosity influence on fiber compaction in tapered resin
injection pultrusion manufacturing. Appl. Compos. Mater. 2018, 25, 485–506. [CrossRef]
15. Jeswani, A.L.; Roux, J.A. Impact of fiber volume fraction and resin viscosity with die-detached tapered
chamber in resin injection pultrusion. J. Manuf. Sci. Eng. 2010, 132, 021007. [CrossRef]
16. Fleischer, J.; Lanza, G.; Mativenga, P.T.; Möhring, H.C.; Caggiano, A. Composite materials parts manufacturing.
In Proceedings of the 1st CIRP Conference on Composite Materials Parts Manufacturing, Karlsruhe, Germany,
8–9 June 2017.
17. Jaklitsch, D.J.; Bostic, M.T.; Pattie, E.R. An Examination of Processing Variables in the Pultrusion of Glass
Reinforcements with an Epoxy-Anhydride Resin System; U.S. Army Materials Technology Laoratory: Watertown,
MA, USA, 1991.
18. Rahatekar, S.S.; Roux, J.A. Epoxy/glass resin injection pultrusion simulation. Polym. Polym. Compos. 2004, 12,
537–550. [CrossRef]
19. Bezerra, R. Modelling and Simulation of the Closed Injection Pultrusion Process. Ph.D. Thesis, Karlsruhe
Institute of Technology, Karlsruhe, Germany, October 2017.
20. Goldsworthy, W.B.; Estates, P.V. Pultrusion Machine and Method. U.S. Patent 3,556,888, 23 June 1967.
21. Strauss, S. Flexibility in polyurethane injection molding. In Proceedings of the North American Pultrusion
Conference 2019, Rosemont, IL, USA, 8–10 April 2019.
22. Gauchel, J.V.; Lehman, R.N. Methods for Resin Impregnated Pultrusion. U.S. Patent 6,048,427, 11 April 2000.
23. Brown, R.J.; Kharchenko, S.; Coffee, H.D.; Huang, I. System for Producing Pultruded Components. U.S.
Patent 8,597,016 B2, 3 December 2013.
24. Krebs, D. Grundlagen der Pultrusion zur Fertigung von FV-Bauteilen für die Anforderungen der automobilen
Großserie. Ph.D. Thesis, Karlsruhe Institute of Technology, Karlsruhe, Germany, March 2018.
25. Koppernaes, C.; Nolet, S.G.; Fanucci, J.P. Method and Apparatus for Wetting Fiber Reinforcements with
Matrix Materials in the Pultrusion Process Using Continuous In-Line Degassing. U.S. Patent 5,073,413,
31 May 1990.
26. Wilhelm, F.; Wiethaler, J.; Karl, R. Power ultrasonic in closed injection pultrusion. In Proceedings of the 18th
European Conference on Composite Materials, Athen, Greece, 24–28 June 2018.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).