Morphology and Mechanical Properties of Layered Silicate

Reinforced Natural and Polyurethane Rubber Blends

Produced by Latex Compounding

S. Varghese,1 K. G. Gatos,2 A. A. Apostolov,3 J. Karger-Kocsis2

1
Rubber Research Institute of India, Kottayam, Kerala-686009, India
2
Institut für Verbundwerkstoffe GmbH (Institute for Composite Materials), Kaiserslautern University of Technology,
P.O. Box 3049, D-67653 Kaiserslautern, Germany
3
Laboratory on Polymers, University of Sofia, BG-1126, Sofia, Bulgaria

Received 5 September 2003; accepted 29 October 2003

ABSTRACT: Natural rubber (NR), polyurethane rubber nanocomposites. It was found that LS is more compatible
(PUR), and NR/PUR-based nanocomposites were produced and thus better intercalated by PUR than by NR. Further, LS
from the related latices by adding a pristine synthetic lay- was preferably located in the PUR phase in the blends,
ered silicate (LS; sodium fluorohectorite) in 10 parts per which exhibited excellent mechanical properties despite the
hundred parts rubber (phr). The dispersion of the LS latices incompatibility between NR and PUR. Nano-reinforcement
in the composite was studied by X-ray diffraction (XRD) and was best reflected in stiffness- and strength-related proper-
transmission electron microscopy (TEM). Further informa- ties of the rubber composites. © 2004 Wiley Periodicals, Inc.
tion on the rubber/LS interaction was received from Fourier J Appl Polym Sci 92: 543–551, 2004
transform infrared spectroscopy (FTIR) and dynamic me-
chanical thermal analysis (DMTA). Tensile and tear tests Key words: clay; latices; nanocomposites; rubber; structure–
were used to characterize the performance of the rubber property relations

INTRODUCTION ous dispersion of fine rubber particles (particle size

below 5 ␮m). Mixing of latex with LS, followed by
Nowadays rubber nanocomposites containing layered
coagulation, is therefore an interesting way to produce
silicates (LS) as reinforcement are gaining impor-
rubber nanocomposites. This route has been already
tance.1 The interest behind this development is due to
followed for natural (NR),4 styrene/butadiene
the nanoscale dispersion (the thickness of the layered
(SBR),5– 6 acrylonitrile/butadiene (NBR),7 and carbox-
silicates is ca. 1 nm) and the very high aspect ratio of
ylated NBR.8 On the other hand, no report is available
the silicate platelets (length-to-thickness ratio up to
on LS-reinforced latex blends. This is quite surprising
2000),2 enabling high reinforcing efficiency even at
as latex combinations are widely used to improve
low LS loading. To make the polar LS compatible with
some praxis-relevant properties of the constituents.
nonpolar polymers and thus to facilitate the exfolia-
Note that NR has to be filled/reinforced owing to its
tion of LS, the silicates are made organophilic
moderate tear strength [e.g., 9 –10]. To improve the
(e.g., 2–3). This occurs by exploiting the cation ex-
resistance to solvents (especially toward hydrocar-
change capacity of the LS. Organophilic LS are, how-
bons), abrasion, and UV irradiation, NR is often
ever, expensive, which forced researchers to have a
blended with polyurethane rubber (PUR). Accord-
look at alternative methods. Nonorganophilic (pris-
ingly, the aims of this present work were (1) to pro-
tine) LS can be dispersed in water, which acts as
duce LS-reinforced NR/PUR-based nanocomposites
swelling agent via hydration of the intergallery cations
via latex compounding, and (2) to study their mor-
(usually Na⫹ ions). Note that several rubbers are
phology-dependent mechanical properties.
available in latex form, which is a rather stable aque-
EXPERIMENTAL
Correspondence to: J. Karger-Kocsis (karger@ivw.uni-kl.de). Materials
Contract grant sponsor: Alexander von Humbolt Founda-
tion. As LS, a synthetic sodium fluorohectorite (Somasif
Contract grant sponsor: German Science Foundation; con- ME-100) of Co-Op Chemicals (Tokyo, Japan) was se-
tract grant number: DFG-GRK 814. lected. This LS had a cation exchange capacity of 100
Contract grant sponsor: German Academic Exchange.
meq/100 g and an intergallery distance of 0.95 nm.
Journal of Applied Polymer Science, Vol. 92, 543–551 (2004) Note that this LS exhibits a very high aspect ratio (viz.
© 2004 Wiley Periodicals, Inc. ⬎1000).4,11
544 VARGHESE ET AL.

TABLE I the LS powder was also registered, however, in reflec-

Formulation of NR Prevulcanized Latex tion.
Formulation TEM images were taken with a LEO 912 Omega
Wet Dry microscope (Oberkochen, Germany) at an accelerator
voltage of 120 kV. Thin sections (ca. 100 nm) of the
NR latex (60%) 166.7 100.0 specimens were cryocut with a diamond knife at ca.
10% KOH solution 1.0 0.1
⫺120°C and used without staining.
50% ZDMC dispersion 2.0 1.0
50% sulfur dispersion 2.0 1.0 To get a deeper insight into the possible interaction
between LS and rubber, Fourier transform infrared
ZDMC, zincdimethyldithiocarbamate. spectroscopic (FTIR) measurements were also done.
FTIR on the films was performed in the attenuated
total reflection mode (ATR) at a resolution of 4 cm⫺1
Sulfur prevulcanized NR latex, along with the in- by using a Nicolet P510 spectrometer (Madison, WI,
gredients, were procured from Rubber Research Insti- USA). LS powder was pressed with KBr powder for
tute of India (Kottayam, Kerala, India). This concen- FTIR measurements in the transmission mode.
trated, high-ammonia (1%) NR latex contained 60%
dry rubber. For prevulcanization, this latex was mixed
with the ingredients listed in Table I under slow stir- Property assessment
ring. The compounded latex was then heated to 70°C
in a water bath with low stirring for 4 h. The prevul- Dynamic mechanical thermal analytic (DMTA) spec-
canized latex obtained was cooled to room tempera- tra of the films were recorded by an Eplexor 25N
ture and the initial ammonia content was restored by device (Gabo Qualimeter, Ahlden, Germany) in ten-
adding ammonia solution. The NR latex was then sion mode at 10 Hz frequency. The complex elastic
stored in tight plastic bottles until use. modulus, its constituents (viz. storage, E⬘, and loss
PUR latex (Impranil DLP-R) containing ⬃ 50% poly- parts, E⬙), along with the mechanical loss factor (tan ␦),
ester-based polyurethane was supplied by Bayer (Le- were determined as a function of the temperature (T
verkusen, Germany). ⫽ ⫺100°C . . . ⫹60°C). The static and dynamic tensile
loads applied were 2 and ⫾1N, respectively, and the
heating rate was set to 2°C/min.
Film casting Tensile tests, to determine the ultimate properties
The prevulcanized NR latex was mixed with the aque- (strength, elongation), along with the moduli at se-
ous dispersion of LS (10%) and stirred well. The dirt lected elongations were performed at room tempera-
and coarse particles were removed by filtering ture (RT) on dumbbells according to ASTM D412 by
through a sieve (with an opening of 250 ␮m) and the using a 500 mm/min crosshead speed. The tear
latex compound was cast in a mold built of glass strength at RT was determined according to ASTM
plates (dimensions: 130 ⫻ 100 ⫻ 2 mm). The casting D624 by using crescent-shaped specimens at a cross-
was allowed to dry in air until transparent and post- head speed of 500 mm/min. The tensile and tear prop-
vulcanized at 100°C for 30 min in an air-circulated erties were determined also after heat aging (storage
oven. Fully vulcanized samples were then cooled and for 7 days at 70°C).
packed in sealed polyethylene bags for testing.
Aqueous dispersion of LS was added to the PUR
latex, stirred, and cast as indicated above and air dried RESULTS AND DISCUSSION
until transparent. Note that the PUR was not cured.
Morphology
Latex blends with various PUR/NR ratios (viz. 1/1
and 8/2) with and without LS were produced in a Figure 1 shows the XRD spectra of the LS and the
similar way as described above. LS-containing films of various compositions. Note that
the LS shows two smaller peaks in addition to the
major peak. These peaks correspond to the following
Morphology detection
interlayer distances based on the Bragg’s equation:
The dispersion of LS in the latex films was studied by 1.22, 1.10, and 0.95 nm; so, the LS used contained some
X-ray diffraction (XRD) and transmission electron mi- small fractions with higher intergallery distance than
croscopy (TEM). XRD spectra were obtained in the the bulk material. LS has been intercalated by NR in
transmission mode by using Ni-filtered CuK␣ radia- the related compound as the interlayer distance of the
tion (␭ ⫽ 0.1542 nm) by a D500 diffractometer (Sie- LS increased to 1.19 –1.31 nm. The appearance of the
mens, Munich, Germany). The samples were scanned related broad peak suggests that the degree of NR
in the step mode at a 1.5o/min rate in the range of 2␪ intercalation is different. Considerably better interca-
⬍ 12o. For comparison purposes, the XRD spectrum of lation was noticed for the PUR latex where two peaks
LAYERED SILICATE REINFORCED RUBBER BLENDS 545

Figure 1 XRD spectra of the layered silicate (LS) reinforced latex nanocomposites of various compositions. [Note. For
comparison purposes, this figure contains the XRD spectrum of the LS (sodium fluorohectorite) as well.]

were resolved. The major peak indicates that the in- the latex. Based on the TEM results, we can now
terlayer distance of the LS widened to 1.73 nm from explain the difference in the XRD spectra of the PUR
the initial 0.95 nm. This effect can be assigned to the and PUR/NR latices. Recall that LS is less intercalated
higher polarity of PUR compared to NR, which favors by NR than by PUR. So, in the case of the PUR/NR
the compatibility with LS. Similar to PUR, the NR/ blend, PUR should intercalate double the amount of
PUR latex blend also shows two peaks. Albeit they LS because the volume is excluded by NR. Bearing in
appear at slightly higher interlayer distances than in mind that there is an optimum in the LS content in
PUR, these peaks are the same. The intensity ratio of respect to intercalation/exfoliation phenomena, a sub-
these peaks is, however, opposed to that of the pure stantial increase in the LS may cause its reaggregation
PUR nanocomposite. Before discussing this aspect, (confinement). However, this does not necessarily
attention should be paid to results achieved by TEM yield a deterioration in the mechanical properties. Re-
and FTIR. call that the prevulcanized NR particles force the sili-
TEM pictures in Figure 2 evidence the good inter- cate aggregates in the neighboring PUR phase to cover
calation of LS by PUR. One may get the impression their surface. This results in a skeleton morphology as
that a part of LS has even been exfoliated. Pictures in the length of the silicate layers is higher than those of
Figure 2 demonstrate further the high-aspect ratio of the diameter of the particles (Fig. 3). The formation of
the LS. This becomes more obvious when the size of this skeleton structure may yield improved mechani-
the flat-laying platelets (disks) in Figure 2(b) is con- cal properties.
sidered. Interesting information can be derived from the
The dispersion of LS in PUR/NR (1/1) latex blend FTIR analysis, too. Several attempts to characterize
differs considerably from that of the PUR. The TEM PUR/clay12–14 or NR/clay15–16 nanocomposites by us-
picture in Figure 3 shows that NR and PUR are not ing FTIR spectroscopy have already been made. In
compatible. Note that particles from the sulfur prevul- most of the cases, just verification of the incorporation
canized NR appear dark in these TEM images. Lay- of the clay into the matrix was the outcome. Differ-
ered silicate stacks can be located at the boundary of ences between the spectra of unfilled material and
the PUR (light) and NR (dark) phases. Pronounced nanocomposite were sought among peaks corre-
intercalation and possible exfoliation took place only sponding to vibrations of the macromolecular chains
in the PUR phase [see Fig. 3(b)]. The silicate layers and of either PUR or NR. Chen et al.12 tried to estimate the
aggregates cover the NR particles, resulting in a skel- degree of interaction between the silicate layers and
eton (house of cards) structure. This peculiar morphol- the PUR segments evaluating the ratio of the absorp-
ogy is rather specific for NR nanocomposites pro- tion peaks of the hydrogen-bonded and the free
duced by the latex route if the length of the LS is groups of NH or CAO. Creation of hydrogen bonds
commensurable with that of the rubber particle size in between functional groups of the polymer matrix and
546 VARGHESE ET AL.

Figure 2 TEM images taken at various magnifications from the film cast of PUR latex containing 10 phr LS.

the organoclay as well as their maintenance on in- monitored the stress-induced peak shift in the Si—O
creasing temperature was examined by Lee and Han stretching vibration of montmorillonite clay in nylon-
for polycarbonate17 and polystyrene-block-hydroxy- 6/nanoclay nanocomposite. The vibration of the
lated polyisoprene copolymer.18 Recently, Loo et al.19 Si—O bond was found to be sensitive to stress, show-
LAYERED SILICATE REINFORCED RUBBER BLENDS 547

Figure 3 TEM pictures taken from the film cast of the PUR/NR (1/1) latex blend containing 10 phr LS.

ing a shift to lower wavenumbers with increasing composition.20 In the case of fluorohectorite, the IR
level of strain. spectrum presents mainly two peaks corresponding to
The absorption bands in the infrared (IR) spectrum the Si—O stretching vibration, ␯ (Si—O), at the 1005
of various layered silicates depend on their chemical cm⫺1, and the Si—O bending vibration, ␦ (Si—O), at
548 VARGHESE ET AL.

Figure 4 FTIR spectra of PUR, LS, and PUR reinforced with LS 10 phr.

476 cm⫺1.13,19,20 The sensitivity of these peaks to in- the layers of LS (i.e., TEM and WAXS experiment), the
tercalation/exfoliation phenomena was observed in peak position is likely to be due to the interaction of
the current article. the macromolecular chain with the silicate layers.
As presented in Figure 4, the Si—O stretching vi- Figure 5 presents the spectra in the case of the
bration, at 1005 cm⫺1 in the case of the PUR/LS sys- NR/LS system. The Si—O stretching vibration, at 1005
tem, appears as a shoulder around 990 cm⫺1 super- cm⫺1, and the Si—O bending vibration, at 476 cm⫺1,
posed on the 967 cm⫺1 peak of PUR. Moreover, the are shifted to 998 and 470 cm⫺1, respectively. Accord-
Si—O bending vibration at 476 cm⫺1 is shifted to 467 ing to the TEM and WAXS findings, the NR/LS sys-
cm⫺1, presenting a clear peak due to the fact that at tem showed less significant intercalation (and thus
that region the PUR does not show any peak. Consid- layer expansion) than PUR. This means that the inter-
ering the fact that the PUR is capable of intercalating action between the NR macromolecular chains and the

Figure 5 FTIR spectra of NR, LS, and NR reinforced with LS 10 phr.

LAYERED SILICATE REINFORCED RUBBER BLENDS 549

Figure 6 FTIR spectra of PUR/NR (1/1) blend, LS, and PUR/NR (1/1) blend reinforced with LS 10 phr.

layered silicate is rather low. Respectively, the peak positions and ratio of the XRD peaks in Figure 1 for
shift in the IR spectra for the NR/clay nanocomposite the NR/PUR blend based composite.
was also smaller than the shift for PUR/clay nano-
composite.
The spectrum of the PUR/NR blend reinforced with
LS is presented in Figure 6. The Si—O stretching vi- Thermomechanical properties
bration, at 1005 cm⫺1, and the Si—O bending vibra- Figure 7 shows the trend of the storage modulus [E⬘,
tion, at 476 cm⫺1, are shifted to 994 and 468 cm⫺1, Fig. 7(a)] and mechanical loss factor [tan ␦, Fig. 7(b)] as
respectively. This means that there is a rather good a function of temperature for the latices studied. Com-
intercalation of LS in the blend, similar to the neat paring the DMTA traces of the plain rubbers with that
PUR. Considering the TEM images, the component of the blend, one can notice that PUR and NR are fully
that worked as an intercalant in the blend was the incompatible. This is based on the fact that no change
PUR rather than the NR. in the related glass transition temperatures (Tg) occurs
Consulting the above-mentioned results, it is clear
due to blending and the stiffness response follows the
that PUR has two favorable peaks in the case of the
composition ratio. This finding is in harmony with the
XRD spectra. This means that there are two favorable
TEM results. The nano-reinforcement proved to be
and possible distances between the layers of the sili-
very efficient below the Tg of the matrix (plain rub-
cate during intercalation. In the case of the spectra
bers) and below the component with the higher Tg
taken from the blend, these two peaks also appear but
(blend rubbers), respectively. The stiffness of the plain
with totally opposite intensity. The volume during
drying the latex compound (glass plates) was the same rubbers was increased by 1200 –1500 MPa (depending
each time and LS is obviously better intercalated by on the temperature) owing to 10% LS. One can notice
PUR than by NR (XRD spectra). Considering the fact that the formation of a skeleton structure in NR and
that, in the blend, the volume of the better intercalat- PUR/NR blend is as efficient as the markedly better
ing PUR is one-half, some excluded volume phenom- intercalation, however, without skeleton structure in
ena may appear, eventually causing restricted mobil- PUR [cf. Figs. 2(a) and 3(a)]. Figure 7(b) demonstrates
ity of the macromolecular chains. The LS is mainly in that nano-reinforcement caused a dramatic decrease
the PUR area (TEM images), so the amount of LS that in the tan ␦. This finding is in agreement with the
should be intercalated by PUR is not actually 10 phr expectation: the molecular mobility is strongly ham-
but almost double. Having in mind that there is an pered owing to the strong LS/rubber interactions.
optimum in LS content for intercalation/exfoliation Note that in Figure 7(a) the major consequence of
processes,1,21–22 the increase of LS content in the PUR blending NR with PUR is obvious: the blend exhibits
should have an adverse effect (i.e., intercalation/exfo- a markedly higher stiffness than NR up to T ⬇ 10°C
liation to a lesser extent). This is reflected by the (Tg of PUR).
550 VARGHESE ET AL.

Figure 7 Storage modulus and mechanical loss factor as a function of temperature for pure and reinforced systems.

Tensile mechanical properties due to their excellent mechanical performance. So,

part of the expensive PUR latex can be replaced by
Table II lists the mechanical properties of the rubbers
inexpensive NR latex without sacrificing the mechan-
and their nanocomposites before and after heat aging.
ical response of the nanocomposites. A further conse-
Note that LS nano-reinforcement was very effective
quence of compounding NR with PUR is related to the
for PUR. The ultimate tensile strength as well as tear
aging of the latter. Heat aging of PUR accompanied by
strength were strongly increased (more than three
crosslinking (via interchange reactions), which en-
times) and a dramatic improvement was found in the
hanced the stiffness and strength data of PUR, PUR-
moduli at different elongations. As expected, the LS
containing blend, and related nanocomposites.
reinforcement reduced the ultimate elongation. A sim-
ilar scenario, however, with less improvement in the
CONCLUSION
stiffness and strength, was found for the NR. The most
interesting results, for the PUR/NR (composition ra- On the basis of this work, devoted to a study of the
tio: 1/1 and 8/2) blend-based nanocomposites, were morphology-dependent mechanical properties of LS
LAYERED SILICATE REINFORCED RUBBER BLENDS 551

TABLE II
Mechanical Properties of the Rubber Nanocomposites Studied
PUR ⫹ LS NR ⫹ LS PUR/NR (1/1) PUR/NR (8/2)
Property PUR 10 phr NR 10 phr ⫹ LS 10 phr ⫹ LS 10 phr

Before aging
Tensile strength (MPa) 4.0 15.9 19.6 23.5 12.4 11.4
Tensile modulus (MPa)
100% Elong. 0.8 5.6 0.7 2.1 4.3 4.9
200% Elong. 0.9 7.8 0.9 3.1 5.9 6.7
300% Elong. 1.1 10.1 1.1 4.5 7.5 8.4
Elongation at break (%) 932 543 881 697 556 469
Tear strength (kN/m) 12.3 54.5 28.0 36.7 59.9 50.7
After aging at 70°C for 7 days
Tensile strength (MPa) 10.5 17.9 20.8 23.5 16.7 17.5
Tensile modulus (MPa)
100% Elong. 1.1 7.6 0.7 2.7 6.7 7.4
200% Elong. 1.4 10.7 0.9 4.2 9.4 10.4
300% Elong. 1.8 13.5 1.1 6.0 11.6 13.0
Elongation at break (%) 772 444 768 620 484 447

reinforced NR-, PUR-, and PUR/NR-blend based 3. Pinnavaia, T. J.; Beall, G. W., Eds. Polymer–Clay Nanocompos-
nanocomposites produced by the latex route, the fol- ites; Wiley: New York, 2000.
4. Varghese, S.; Karger-Kocsis, J. Polymer 2003, 44, 4921.
lowing conclusions can be drawn.
5. Zhang, L.; Wang, Y.; Wang, Y.; Sui, Y.; Yu, D. J Appl Polym Sci
LS is more compatible and thus better intercalated 2000, 78, 1873.
by PUR than by NR. In the case of sulfur-prevulca- 6. Wang, Y.; Zhang, L.; Tang, C.; Yu, D. J Appl Polym Sci 2000, 78,
nized NR latex, and its blends with PUR, the LS forms 1879.
a skeleton (house of cards) structure. The onset of this 7. Wu, Y.-P.; Jia, Q.-X.; Yu, D.-S.; Zhang, L.-Q. J Appl Polym Sci
structure is favored by the prevulcanization of the NR. 2003, 89, 3855.
The reinforcing efficiency of the skeleton-type struc- 8. Wu, Y.-P.; Zhang, L.-Q.; Wang, Y.-Q.; Liang, Y.; Yu, D.-S. J Appl
Polym Sci 2001, 82, 2842.
ture (NR) was comparable with that composed of LS
9. Cai, H.-H.; Li, S.-D.; Tian, G.-R.; Wang, H.-B.; Wang, J.-H. J Appl
layers and stacks (PUR).
Polym Sci 2002, 87, 982.
Albeit that PUR and NR are completely incompati-
10. Stephen, R.; Raju, K. V. S. N.; Nair, S. V.; Varghese, S.; Oommen,
ble, the mechanical properties of the nanocomposites Z.; Thomas, S. J. Appl Polym Sci 2003, 88, 2639.
based on their blends (PUR/NR ratios 1/1 and 8/2) 11. Varghese, S.; Karger-Kocsis, J.; Gatos, K. G. Polymer 2003, 44,
agreed with those of the plain PUR. The effect of LS 3977.
dispersion (intercalation/exfoliation) was best re- 12. Chen, T. K.; Tien, Y.-I.; Wei, K.-H. Polymer 2000, 41, 1345.
flected in stiffness- and strength-related characteris- 13. Zhang, X.; Xu, R.; Wu, Z.; Zhou, C. Polym Int 2003, 52, 790.
tics. 14. Tien, Y. I.; Wei, K.-H. Polymer 2001, 42, 3213.
15. Joly, S.; Garnaud, G.; Ollitrault, R.; Bokobza, L.; Mark, J. E.
Dr. S. Varghese, K. G. Gatos, and Dr. A. A. Apostolov Chem Mater 2002, 14, 4202.
acknowledge the support of the Alexander von Humbolt 16. Arroyo, M.; López-Manchado, M. A.; Herrero, B. Polymer 2003,
Foundation, German Science Foundation (DFG-GRK 814), 44, 2447.
and German Academic Exchange (DAAD), respectively. 17. Lee, K. M.; Han, C. D. Polymer 2003, 44, 4573.
18. Lee, K. M.; Han, C. D. Macromolecules 2003, 36, 804.
19. Loo, L. S.; Gleason, K. K. Macromolecules 2003, 36, 2587.
References 20. Mookherjee, M.; Redfern, S. A. T. Clay Miner 2002, 37, 309.
1. Karger-Kocsis, J.; Wu, C.-M. Polym Eng Sci to appear. 21. Alexandre, M.; Dubois, P. Mater Sci Eng Reports 2000, 28, 1.
2. Utracki, L. A.; Kamal, M. R. Arab J Sci Eng 2002, 27, 43. 22. Kim, J.-T.; Oh, T.-S.; Lee, D.-H. Polym Int 2003, 52, 1058.