See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/230192993

Effect of glass fibers on rheology, thermal and mechanical properties of

recycled PET

Article in Polymer Composites · March 2008

DOI: 10.1002/pc.20467

CITATIONS READS

36 1,942

7 authors, including:

Milan Kracalik Ladislav Pospíšil

Johannes Kepler University of Linz Masaryk University
54 PUBLICATIONS 502 CITATIONS 6 PUBLICATIONS 94 CITATIONS

SEE PROFILE SEE PROFILE

Miroslav Slouf Jana Mikesova

Czech Academy of Sciences, Institute of Macromolecular Chemistry Czech Academy of Sciences, Institute of Macromolecular Chemistry
479 PUBLICATIONS 8,877 CITATIONS 45 PUBLICATIONS 591 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Milan Kracalik on 21 December 2017.

The user has requested enhancement of the downloaded file.

Effect of Glass Fibers on Rheology, Thermal and
Mechanical Properties of Recycled PET

Milan Kráčalı́k,1 Ladislav Pospı́šil,2 Miroslav Šlouf,3 Jana Mikešová,3 Antonı́n Sikora,3,4
Josef Šimonı́k,4 Ivan Fortelný3,4
1
Institute of Plastics Processing, University of Leoben, Franz-Josef Str. 18, 8700 Leoben, Austria

2
Polymer Institute Brno Ltd., Tkalcovská 2, 656 49 Brno, Czech Republic

3
Institute of Macromolecular Chemistry, Academy of Sciences of the Czech Republic,
Heyrovský Sq. 2, 162 06 Prague, Czech Republic

4
Faculty of Technology, Tomas Bata University in Zlı́n, T.G.M. Sq. 275, 762 72 Zlı́n, Czech Republic

PET-glass fiber composites were prepared by melt mix- PET-reprocessing countries varies usually around 20 wt%.
ing of recycled PET with chopped glass fibers (15, 20, The rest of the used bottles are incinerated or dumped [2,
and 30 wt%) and their degree of dispersion was 3]. A further problem consists in limited applications of
assessed by scanning electron microscopy. Rotational
rheometry was employed to analyze the interfacial recycled PET (mostly staple fibers, packaging, films, and
shear strength between the fibers and polymer matrix strips) [2]. The limitation is due to processing properties
in the molten state. The composite containing 30 wt% of the recyclate (viscosity, melt strength), which are deter-
of glass fibers revealed a moderate G9 secondary pla- iorated during recycling process. This problem can be
teau; hence strong fiber-matrix interactions were con- solved by incorporation of proper particles into recycled
firmed. Results of mechanical testing were in a good
accordance with structural and rheological measure- polymer by melt compounding. Consequently, melt vis-
ments. The higher rate of mixing under production- cosity increases in the range of low shear rates and
scale conditions resulted in lower fiber-matrix adhe- enhanced melt strength makes reprocessing possible. The
sion and in a similar level of fiber dispersion as com- use of glass fibers involves interesting combination of
pared to the same mixture compounded on pilot-plant high reinforcement with a rather low price. However,
scale. Thermal characterization of the composites was
performed by differential scanning calorimetry and total enhancement of processing and utility properties of
crystalline fraction was analyzed. POLYM. COMPOS., recycled polymers requires high interfacial adhesion
29:915–921, 2008. ª 2008 Society of Plastics Engineers between glass fiber and polymer matrix. Sufficient fiber-
matrix interactions are essential to transfer the stresses
from the matrix to the reinforcing fibers. Therefore, analy-
INTRODUCTION sis of interfacial shear strength from rheological measure-
ments in the linear viscoelasticity region gives necessary
Preparation of composites using recycled polymers is
information on the effect of reinforcing particles on pro-
an important issue both from the technological and envi-
cessing behavior. Moreover, material characteristics in the
ronmental points of view. The amount of poly(ethylene
melt state reflect structural changes, which have a strong
terephthalate) (PET) used in plastics industry (especially
impact on the final performance of dispersive polymeric
for beverage bottles) has been significantly increasing [1],
system. The recycled PET-glass fiber composites have
hence the problem of sufficient recycling is more and
been studied so far; however, no study dealing with pro-
more challenging. The volume of recycled PET bottles in
cessing properties (exact analysis of elastic and viscous
part of melt viscosity) and their correlation with structure
Correspondence to: M. Kráčalı́k; e-mail: [email protected] and utility performance has been found in literature. The
Contract grant sponsor: Ministry of Industry and Trade of the Czech
current papers focus mostly on thermal and mechanical
Republic, Polymer Institute Brno, Ltd; contract grant number: 1H-PK2/67.
DOI 10.1002/pc.20467
(possibly flammability) characterization [4–10] and rarely
Published online in Wiley InterScience (www.interscience.wiley.com). on degradation analysis or processing/structural changes
V
C 2008 Society of Plastics Engineers in recycled PET-glass fiber systems [11, 12]. Neverthe-

POLYMER COMPOSITES—-2008
less, understanding rheological properties and structural Scanning Electron Microscopy
changes and, consequently, material end-properties
The scanning electron microscopy (SEM) micrographs
belongs to the needs of applied polymeric engineering.
were obtained with JSM 6400 (Jeol). The specimens for
SEM observations were prepared in two ways. The first
EXPERIMENTAL series of specimens was cut with microtome at room tem-
perature and the other series was fractured in liquid nitro-
Materials gen. The cut and fracture surfaces of the specimens from
the first and second series were examined by SEM mi-
Color-sorted recycled poly(ethylene terephthalate) croscopy. Before microscopic investigations, all speci-
(PET-R), with the intrinsic viscosity 0.93 dl/g (phenol/tet- mens were sputtered with thin platinum layer (ca. 4 nm)
rachloroethane 1:3) as well as ECTA ES2001EF chopped in a vacuum sputter coater SCD 050 (Balzers) to prevent
glass fibers (original fiber length 4 mm, diameter 10 mm) charging and decrease the sample damage by the electron
were supplied by Polymer Institute Brno. According to beam. All micrographs were secondary electron images
characteristics in Table 2, PET-R exhibits usual thermal taken at an acceleration voltage 30 V.
behavior (similar to virgin bottle grade PET) because no re-
markable decrease in Tg (Tm) magnitude occurred indicating
quite a low level of impurity. To improve processing and Thermal Characteristics
utility properties (higher interfacial adhesion, dispersion, Thermal characterization of the polymer matrix and
stability of the mixture), 10% of Naintsch A-3 talc, 1% of nanocomposites was carried out by differential scanning
Licowax OP flow modifier, and 0.2% of Ultranox 626 sta- calorimetry (Perkin Elmer, Pyris 1 DSC) using the stand-
bilizer were admixed before compounding (Table 1). ard mode: 1, 308C for 3 min; 2, heating from 308C to
2808C at 108C/min; 3, 2808C for 2 min. The thermal pa-
Preparation of Composites rameters, glass transition temperature (Tg), cold crystalliza-
tion temperature (Tc), melting temperature (Tm), enthalpy
The recycled PET flakes were dried at 1208C in an of cold crystallization (DHc), and enthalpy of melting
oven for 6 h. The polymer was mixed with glass fibers, (DHm) were calculated for the matrix and polymer fraction
talc, flow agent, and stabilizer in a corotating twin-screw in the composites. The crystalline content (Xc) was eval-
extruder (Werner & Pfleiderer ZSK 25, d ¼ 25 mm, pilot- uated by assuming the DHm of hypothetical 100% crystal-
scale output 10 kg/h; for comparison, production-scale line poly(ethylene terephthalate) to be 117.6 J/g [13].
output conditions at 250–400 kg/h were tested). The com-
pounding temperature was 260–2758C and the screw
speed 200 rpm. For characterization, recycled PET com- Mechanical Properties
posites were injection-molded (Battenfeld BA750/200) to
For tensile tests an Instron 5800 R was employed.
specimens at 260–2758C, (mould temperature 40–1208C).
Experiments were performed according to the ISO 527-2
standard. The crosshead speeds were 1 mm/min for tensile
Melt Rheology modulus measurements and for all other characteristics
50 mm/min. The toughness characterized as Charpy impact
Rheological properties were studied using an ARES 3
strength was measured according to the CSN EN ISO 179-1
Rheometer (Advanced Rheometric Expanded System, Rheo-
standard at room temperature.
metric Scientific) with parallel-plate geometry and 25-mm
diameter plates. All measurements were performed with
two automatically switched force transducers with a torque RESULTS AND DISCUSSION
range of 0.02–2,000 g cm. The sample thickness ranged
from 0.9 to 1.1 mm. Experiments were performed at 2708C Morphology of Glass Fibers in Recycled PET
under nitrogen to prevent degradation of samples. The fol-
lowing types of rheological measurements were carried out: The prepared composites were analyzed by scanning
(1) dynamic strain sweep test (at 6.28 rad/s) to confirm the electron microscopy in the solid state to determine
linearity of viscoelastic region, (2) dynamic frequency changes in the structure caused by the introduction of
sweep test over a frequency range of 0.1–100 rad/s, at the glass fibers into recycled PET. The effects of variations in
strain 2% for the composites, 30% for the matrix. fiber concentration and compounding conditions on the
structure of polymer matrix were investigated and the
TABLE 1. Specification of mixtures.
results are shown in Figs. 1–3.
The influence of the fiber content and processing con-
Composition (wt%) W 5167 W 6187 W 5455 GF20T10 ditions on the structural changes is indicated in Figs. 1
and 2. With increasing fiber concentration, the homogene-
PET-R 73.8 68.8 58.8 68.8 ity of fiber distribution in polymer matrix remains almost
ECTA ES2001EF 15 20 30 20
constant. The remarkable change in structure was detected

916 POLYMER COMPOSITES—-2008 DOI 10.1002/pc

 FIG. 1. SEM micrographs of fracture surfaces of composites.

in the GF20T10 system, where air bubbles were observed be seen in Fig. 4, the recycled PET matrix showed New-
in both fracture (Fig. 1) and cut surfaces (Fig. 2). The av- tonian behavior in the frequency dependence Z*(o) within
erage diameter of the bubbles varies around 30 mm so the investigated range of shear rates.
that confusion of air bubbles with fiber cavities can be Compared with the unfilled matrix, a significant growth
excluded. This structural inhomogeneity (remarkable in of the complex viscosity of composites in the range of
Fig. 2) is attributed to higher processing speed under pro- low frequencies (more than one order) occurred. All the
duction-scale conditions. prepared composites show a shear thinning phenomenon,
According to Fig. 3, a similar level of fiber-matrix which is caused by orientation of fibers in the flow direc-
interfacial adhesion in all composite systems was tion.
achieved. It is obvious from our previous work [14] that The highest filling effect was achieved with 30% load-
talc particles in PET/glass fibers composite systems act as ing of glass fibers. A lower content of the filler in the
a coupling agent and, consequently, facilitates interfacial polymer matrix manifested itself by lower melt viscosity
interactions resulting in high interfacial shear strength. in the whole range of shear rates. The flow curves of the
W 6187 and GF20T10 systems revealed similar tendency,
whereas the composite prepared under production-scale
conditions showed slightly lower melt viscosity.
Melt Rheology
The internal structural changes in composite systems
Dynamic rheological properties of PET-R/glass fiber under shear flow can be analyzed using frequency
composites were investigated in the region of linear vis- dependences of the storage (G0 ) and loss (G00 ) moduli.
coelasticity. The dynamic strain sweep test [G0 (g)] Addition of a filler to the PET-R melt causes an increase
revealed the linearity region ranging at 1–100% strain for in the dynamic moduli, particularly in G0 (Fig. 5). The
the matrix and 1–15% strain for the composites. As can unfilled matrix behaves like viscoelastic liquid (G00 > G0 ).

DOI 10.1002/pc POLYMER COMPOSITES—-2008 917

 FIG. 2. SEM micrographs of cut surfaces of composites.

The higher value of G0 than G00 for the composites shows Moreover, the flow and melt strength curves are in
a change in rheological behavior, the ‘‘liquid-solid’’ trans- good agreement with structure determined by scanning
formation. Furthermore, the power-law dependence of the electron microscopy (a higher fiber content in polymer
dynamic moduli at low frequencies, which is typical of matrix results in higher melt strength and viscosity of
neat PET, is absent in the composites. The dependence of composites, while a lower fiber-matrix interfacial adhe-
G0 (o) at low shear rates is nearly invariable. This indica- sion together with porosity of GF20T10 system manifests
tion of a G0 (o) secondary plateau can be associated with itself by worse rheological properties compared with the
similar behavior of well-dispersed structures in polymer corresponding W 6187 composite).
nanocomposites, where high interfacial shear strength is
parallel with the formation of network structure (exfolia-
Thermal Characteristics
tion) of silicate layers [15–18].
Usually, fiber-filled polymer dispersions exhibit a fre- Thermal properties of composites were studied by dif-
quency dependence of storage modulus similar to matrix ferential scanning calorimetry. The values of thermal pa-
increasing with the filler content [19]. The composites rameters are listed in Table 2. A moderate increase in
prepared in our study showed a moderate G0 secondary glass transition temperature with glass fibers loading was
plateau and, thereby, confirmed high fiber-matrix adhe- observed. This phenomenon can be explained by strong
sion in the molten state, whereas number of surface polymer-fiber interactions, which cause higher fixation of
interactions rise with glass fiber loading. This increase in polymer chains surrounded by glass fibers. This explana-
melt strength facilitates further processing of material in tion of high interfacial adhesion is in a good agreement
such cases, where the loss of melt cohesion causes pro- with results of other testing methods.
cess difficulties (e.g. during flat-die extrusion or blow In comparison to the GF20T10 system (Fig. 6), all the
moulding). composites from the W series exhibited unusual thermal

918 POLYMER COMPOSITES—-2008 DOI 10.1002/pc

 FIG. 3. SEM micrographs of fracture surfaces of composites at higher magnification.

behavior characterized by first endothermic melting peak processing history where crystallization on the fiber sur-
Tmf at 183–1928C and by the absence of exothermic peak face occurred during lab-scale extrusion. The amount of
of cold crystallization. This difference reflects variation in first crystalline fraction decreased with fiber loading,

FIG. 4. The complex viscosity vs. frequency for the recycled PET ma- FIG. 5. The storage modulus vs. frequency for the recycled PET matrix
trix and composites. and composites.

DOI 10.1002/pc POLYMER COMPOSITES—-2008 919

 TABLE 2. Thermal properties of composites and neat matrix. TABLE 3. Mechanical properties of the composites.

Tga Tmfb DHmfc Tmd DHme Xcf Properties W 5167 W 6187 W 5455 GF20T10
Composite (8C) (8C) (J/g) (8C) (J/g) (%)
Tensile strength (MPa) 110 113 121.7 111
W 5167 74.0 183.4 3.1 252.9 36.9 31.4 sx 1.2 1.3 1.5 1.1
W 6187 75.2 187.8 2.69 253.7 29.4 25 Elongation at break (%) 2.6 2.2 2.2 2.4
W 5455 76.8 191.9 1.7 247.5 25.3 21.5 sx 0.1 0.1 0.1 0.1
GF20T10 75.6 – – 253.2 32.3 27.5 Tensile modulus (MPa) 8,200 10,290 13,900 10,111
PET-R 75.6 – – 252.2 43.2 36.7 sx 52 162 170 138
Flexural modulus (MPa) 7,900 9,377 12,650 9,204
a
Glass transition temperature. sx 47 57 151 48
b
First endothermic melting peak. Impact strength (kJ/m2) 32 43.3 36.6 37.0
c
Enthalpy of first melting (first crystalline fraction). sx 3.1 4.2 3.7 3.5
d
Melting point. Notch impact strength (kJ/m2) 4.6 6.6 6.6 4.8
e
Enthalpy of melting. sx 2.1 3.3 3.3 32.4
f
Crystalline fraction content.

which can be associated with increase in melt shear rate not significantly influenced by different content of added
resulting in higher degradation of macromolecular chains; fibers. The highest stiffness, tensile strength, and flexural
consequently, the number of spherolites during heteroge- modulus exhibited the system filled with 30% of glass
neous crystallization is restricted. This explanation is sup- fibers.
ported by the lowest melt temperature and total crystallin-
ity of W 5455 composite containing the highest loading CONCLUSIONS
of glass fibers. The relative crystalline content of all the
composites was lower compared with that of the neat Recycled PET-glass fiber composites were prepared by
PET-R matrix. melt compounding. According to analysis of SEM micro-
graphs, all the composite systems revealed a high level of
fiber dispersion and homogeneity in the recycled PET ma-
Mechanical Characteristics trix. The highest level of fiber-matrix interfacial adhesion
was obtained with the fiber filling 20%. According to
Mechanical properties of composites are listed in Table 3. rheological study, the highest melt strength was achieved
The values of tensile strength, stiffness, flexural modulus, at the 30% loading of fibers. A significant increase in the
and toughness increased with fiber loading. The process- complex viscosity and storage modulus with glass fibers
ing under production-scale conditions (employed for and talc loading was observed at low frequencies, where
GF20T10 composite preparation) led to slight deteriora- the viscoelastic liquid of recycled PET changed into a
tion of mechanical properties in comparison to analogous solid-like behaviour. In all the mixtures, filling with fibers
system W 6187. However, W 6187 and GF20T10 systems and talc exhibited an enhancing effect on rheological
revealed the highest toughness, probably due to optimal properties of the composites. On the contrary, processing
concentration of glass fibers (see chapter Morphology) under production-scale conditions led to formation of air
and coupling effect of talc particles. The extensibility was bubbles together with the moderate loss of melt strength
and viscosity, compared with analogous system processed
on the pilot-plant scale. A correlation between the fiber
content on linear viscoelastic flow characteristics and dis-
persion level (SEM) of the prepared composites was
found. Mechanical characterization of composites filled
with 20% of glass fibers revealed the highest toughness,
while the 30% fiber loading resulted in the highest level
of stiffness, tensile strength and flexural modulus. Ther-
mal characterization of composites revealed an increase in
glass transition temperature and decrease in total crystal-
linity with fibers loading. Unusual thermal behaviour of
composites prepared under lab-scale conditions was attrib-
uted to crystallization of PET chains on the surface of
glass fibers.

REFERENCES

FIG. 6. DSC thermogram of composites filled with 20 wt% of glass 1. PETCORE Association (PET containers recycling Europe);
fibers. http://www.petcore.org.

920 POLYMER COMPOSITES—-2008 DOI 10.1002/pc

 2. PET Recycling Server, http://www.petrecycling.cz. 11. M.P. Foulc, A. Bergeret, L. Ferry, P. Ienny, and A. Crespy,
3. M. Kráčalı́k, L. Pospı́šil, J. Šimonı́k, J. Kimmer, and J. Polym. Degrad. Stab., 89, 461 (2005).
Hrnčiřı́k, Plasty Kauc., 40, 356 (2003). 12. R. Assadi, X. Colin, and J. Verdu, Polymer, 45, 4403 (2004).
4. A.L.F. de M. Giraldi, J.R. Bartoli, J.I. Velasco, and L.H.I. 13. A. Metha and B. Wunderlich, J. Polym. Sci.: Polym. Phys.
Mei, Polym. Test., 24, 507 (2005). Ed., 16, 289 (1978).
5. A.L.F. de M. Giraldi, R. Cardoso de Jesus, and L.H.I. Mei, 14. M. Kráčalı́k, L. Pospı́šil, M. Šlouf, J. Mikešová, A. Sikora,
J. Mater. Process. Technol., 162/163, 90 (2005). J. Šimonı́k, and I. Fortelný, Polym. Compos., accepted for
6. A. Pegoretti and A. Penati, Polymer, 45, 7995 (2004). publication.
7. A. Pegoretti and A. Penati, Polym. Degrad. Stab., 86, 233 15. R. Krishnamoorti and E.P. Giannelis, Macromolecules, 30,
(2004). 4097 (1997).
8. M.L. Maspoch, H.E. Ferrando, D. Vega, A. Gordillo, J.I. 16. S.A. Khan and R.K. Prudhomme, Rev. Chem. Eng., 4, 205 (1987).
Velasco, and A.B. Martı́nez, Macromol. Symp., 221, 175 17. R. Krishnamoorti, R.A. Vaia, and E.P. Giannelis, Chem.
(2005). Mater., 8, 1728 (1996).
9. F. Laoutid, L. Ferry, J.M. Lopez-Cuesta, and A. Crespy, 18. M. Kráčalı́k, J. Mikešová, R. Puffr, J. Baldrian, R.
Polym. Degrad. Stab., 82, 357 (2003). Thomann, and C. Friedrich, Polym. Bull., 58, 313 (2007).
10. K. Tóth, T. Czvikovszky, and M. Abd-Elhamid, Radiat. 19. R. Guo, J. Azaiez, and C. Celine Bellehumeur, Polym. Eng.
Phys. Chem., 69, 143 (2004). Sci., 45, 385 (2005).

DOI 10.1002/pc POLYMER COMPOSITES—-2008 921

View publication stats