1. R. Gelling MALAYSIAN RUBBER PRODUCERS' RESEARCH ASSOCIATION, TUN ABDUL RAZAK LABORATORY, BRICKENDONBURY, HERTFORD, SG13 8NL, ENGLAND
MODIFICATION OF NATURAL RUBBER LATEX WITH PERACETIC ACID*
MALAYSIAN RUBBER PRODUCERS' RESEARCH ASSOCIATION, TUN ABDUL
RAZAK LABORATORY, BRICKENDONBURY, HERTFORD, SG13 8NL, ENGLAND
The reaction of natural rubber (NR) and other unsaturated polymers with epoxidation reagents is well documented in the literature. However, there are fewer data available on the properties of the products, and in some cases the results are conflicting. NR was first epoxidized in 1922 1 , and the solution epoxidation of other unsaturated polymers has been reviewed by Greenspan 2 . The epoxidation of polymer latices with a variety of reagents has also been reported
The chemistry of epoxidation of unsaturated compounds and subsequent ring-opening reactions have been extensively studied 6 7 . Epoxidation is a stereospecific process8 , and its rate is governed by the substituents on the double bond 9. The ease and position of ring opening of epoxides is again controlled by neighboring groups; both electronic and steric factors are important10
Acidity and temperature appear to be the main factors that govern the nature of the product from the reaction of NR with peracids. At high acid strength and or temperature, the major products are derived from secondary ring-opening reactions However, epoxidized NR (ENR) can be obtained under milder
Udipi reported lS that unvulcanized epoxidized styrene-butadiene block copolymers have tensile strengths and resistance to ASTM oils comparable to vulcanized polychloroprene and nitrile rubbers. It has also been claimed S that epoxidation of NR and other unsaturated elastomers increases wear and solvent resistance and improves the tensile strength and other mechanical properties. However, peroxide cured epoxidized NR vulcanizates were found to have lower tensile strengths compared to an NR control vulcanizate 16 . Roux l ' et al. studied the solution epoxidation of cis-l,4-polyisoprene and reported on the sulfur vulcanizate properties at two epoxidation levels. The tensile properties of the 10 mole% epoxidized material were comparable to those of the
unmodified polymer, but at 30 mole% epoxidation a large reduction in tensile strength and elongation at break was observed, and this was associated with an increase in hardness.
The different physical properties, especially with the latter study, could be attributed to the presence of secondary ring-opened structures. Consequently, the epoxidation of NR with peracetic acid solution has been studied over a range of reaction conditons, and the products characterized, and their properties evaluated.
The NR latex employed in this study was a low ammonia TMTD/ZnO (LATZ) Malaysian stabilized concentrate. Peracetic acid was used as a 35 w/w% so•Received June 25, 1984. B.
lution in acetic acid (Interox Chemicals Ltd.) and was free from strong acid. The nonionic surfactant was an alcohol ethylene oxide condensate (Texofor A30, ABM Chemicals Ltd.).
25 mole% epoxidized natural rubber (ENR-25).—LATZ latex concentrate
(400 g), dry rubber content (d.r.c.) 60 w/w% was stabilized with Texofor A30 (3 phr) and diluted to 960 cm 3 with distilled water. Peracetic acid solution (193 g) was added slowly to the stirred latex over 3 h at IOOC. The latex was stirred for a further 6 h with the temperature maintained below 200C and then coagulated by the addition of methanol. The coagulum was washed with water on a creping mill and soaked in 0.1 M sodium carbonate solution for 16 h to ensure complete removal of acid. The rubber was again washed with water and finally dried in air at 50-600C. A formula of (C20H320)n was obtained by elemental analysis after correction for the oxygen content (1.6%) of the unmodified N R. In the infrared spectrum there were absorptions at 870 and 1240 cm- (epoxide), but no carbonyl or hydroxyl absorptions were visible [ I H NMR (CDC13, ppm) 1.28 (CH3-CO), 2.70 (CH-C-O) 1.65 (CH3-C = C), 5.05 (CH = C)]. An epoxidation level of 26 mole% was obtained by measuring the peak areas of the olefinic and epoxide protons.
ENR-50 and ENR-75.—These were prepared by the above procedure except that the d.r.c. of the latex prior to addition of the peracid was reduced to 20 and 15% respectively and the amount of peracid was increased proportionately. Elemental analysis and I H NMR confirmed the expected epoxide level and together with an infrared spectrum showed the absence of ring-opened structures.
Ring-opened products.—At higher acid concentration and temperatures, complex products containing ring-opened structures were obtained. A repeat of the ENR-50 preparation, but with a d.r.c. of 33% prior to addition of the peracid and no temperature control, resulted in a rubber roduct whose IR spectrum showed additional absorptions at 36 000-3200 cm- (hydroxyl), 1728 cm -I (carbonyl) and 1068 cm-I (cyclic ether). The I H NMR spectrum also had additional signals at 3.8, 3.75, 3.25, 1.10, 1.12, and 1.4 ppm, and their total peak areas indicated that 15 mole% of the original epoxide groups had undergone ring opening.
A similar reaction, but with sufficient peracid for 100 mole% epoxidation, yielded a white amorphous powder of formula (C25H4207)n from elemental analysis [IR spectrum 3600-3200 cm-I (hydroxyl), 1068 cm-I (cyclic ether)]. The I H NMR spectrum was of poor quality, brs. 3.0-4.5 ppm and 1.0-2.5 ppm but no signals due to NR or ENR were observed.
Vulcanizates were prepared by conventional mixing and vulcanization techniques employing the formulations recorded in Table I, All the rubbers were cured to optimum at 1500C and physical properties determined according to the appropriate British Standard, except where indicated. Rebound resilience was measured by dropping a 1.5 mm diameter steel ball from a height of 200 mm under vacuum onto a 1.5 mm thick disc of rubber bonded to a copper block containing a thermocouple. The sample was heated from —100 to +1000C at approximately I OC/min. Skid resistance measurements were carried out on a pendulum apparatus (l m s-l ) designed by the Road Research Laboratory. Stress relaxation was measured at 100% strain at 300 C.
RESULTS AND DISCUSSION
Sufficient peracetic acid was employed to achieve epoxidation levels from 25 to 100 mole% and the effect of concentration and reaction temperature on the structure and properties of the products investigated. Two distinct types of product were isolated: at high reagent concentrations and temperatures, epoxide ringopened products predominated, whereas at low acid concentrations and moderate temperatures, epoxidized NR was isolated as the sole product, except in the case of 100 mole% reaction.
Samples of 25, 50, and 75 mole% (ENR) were prepared by the technique described in the experimental section.
Within the limits of detection of the analytical techniques employed, no ringopened structures could be detected. IR spectra showed absorptions at 870 and 1240 cm-I (epoxide), but no absorptions due to hydroxyl or carbonyl groups which are characteristic of epoxide ring-opened structures. The I H nuclear magnetic resonance (NMR) spectra were consistent with published data on epoxidized synthetic The signals at 2.7 and 5.05 ppm were used to determine the degree of epoxidation, and good agreement was observed between these results and elemental oxygen data.
The properties of these materials will depend to a significant extent on the epoxide unit sequence distribution. Hayashi et al, employed 13 C NMR to determine the epoxide sequence in partially epoxidized cis-l,4-polybutadiene 19 and In both
these solution-prepared materials, a random distribution of epoxide units along the polymer backbone was observed. The present materials were prepared from latex, and the physical constraints of this heterogeneous system could well control the epoxide unit distribution. The 13C
NMR spectra of 20 (Figure l) and 50 mole% epoxidized NR were recorded in deuterochloroform and the signals assigned as shown below:
The sequence distribution of epoxide units is consistent with that expected from a totally random epoxidation process.
The glass transition temperature (Tg) of ENR's, as measured by differential scanning calorimetry, were single sharp events and the value increased with increasing level of
epoxidation (Figure 2). Epoxide ring-opened structures result in a broadening of the Tg and in their absence, this technique can be used to determine the epoxide level.
Gel permeation chromatography (GPC) of ENR's in tetrahydrofuran (Figure 3) indicates that the molecular weights decrease with increasing levels of epoxidation. However, the position is complicated by the fact that the solubility decreases with increasing level of epoxidation due to higher gel contents.
ENR can be crosslinked using any of the standard sulfur formulations normally employed for NR or by a peroxide, although in the latter case the efficiency decreases at epoxide levels of over 50 mole%. The vulcanization characteristics of ENR-25 and ENR-50 (25 and 50 mole% epoxidized respectively) in a semi-EV system (Formulation l , Table I) are compared to those of an NR control
in Figure 4. The main effect of epoxidation is to decrease the scorch delay and hence shorten the cure time. The vulcanization chemistry of ENR will be discussed in detail in a subsequent publication21
Vulcanizates have been prepared from 25, 50, and 75 mole% epoxidized NR and their physical properties determined and compared to those of an NR control (Table Il). The tensile strengths of these ENR gum vulcanizates are high and characteristic of polymers that undergo strain crystallization. An x-ray diffraction study22 has confirmed that ENR does indeed undergo strain crystallization. This is not altogether unsurprising, as the epoxidation process is stereospecific thus maintaining the cis-l,4•configuration of the polymer where the relatively small Qxygen atom can fit into the NR crystal lattice without undue strain. Beyond 50 mole% epoxidation there is a distinct reduction in the degree of crystallinity22, and this is reflected in the nonrelaxing fatigue data in Table Il. How-
20 40 t(min) at 150 0C
FIG. 4.—Monsanto rheographs at 1500C
ever, the fatigue life of ENR-75 is still over an order of magnitude greater than that of a comparable noncrystallizing NBR vulcanizate (24 kHz), The ability to strain-crystallize is also observed in the effect of rate on the tear strength of these polymers (Figure 5):
The increase in the Tg of ENR with degree of epoxidation is reflected in a reduction in room temperature resilience and in other properties. The change in rebound resilience is illustrated in Figure 6. The dramatic improvement in Akron abrasion resistance observed with increasing levels of epoxidation (Table Il) is probably due at least in part to the increase in hysteresis. Schallamach 23 has shown that hysteresis is an important factor in the abrasion testing of rubbers using slipping wheels.
The change in hysteresis also results in improved wet grip performance, but
Formulation I from Table I except where otherwise noted, Formulation Il from Table
less good traction on ice (Table Ill). The optimum combination of wet and ice skid resistance appears to occur at around the 25 mole% epoxidation level.
Epoxidation of NR improves its resistance to hydrocarbon oils and reduces its gas permeability. Increasing levels of epoxidation significantly improve the resistance to hydrocarbon oils and solvents, although the reverse is true for more polar liquids such as hydraulic brake fluids (Table IV). The air permeabilities of ENR•25, -50 and -75 are compared to some synthetic polymers in Table V. At epoxidation levels of over 50 mole%, a ten-fold decrease in permeability was observed, and these materials have similar permeability properties to IIR and NBR (41% acrylonitrile).
With increasing levels of epoxidation, the degree of unsaturation in these materials decreases, and this, as could be expected, results in a reduction in the
rate of oxygen absorption (Figure 7). However, the air aging of a conventional sulfur ENR-50 vulcanizate (Formulation Ill from Table I) was found to be inferior to that of the corresponding NR compound. A large increase in M300 (+230%) and associated decrease in elongation at break (60%) was observed after one day at 1000 C in air. As the ratio of sulfur to accelerator in the formulation decreases towards an EV type system, the aging characteristics of ENR become comparable to those of NR. The aging mechanism of ENR will be discussed in detail in a subsequent publication21
As the acid concentration and temperature of the epoxidation reaction are raised, some of the epoxide groups undergo secondary ring-opening reactions and this is reflected in the properties of the materials. An ENR-50 vulcanizate in which approximately 15% of the epoxide groups had undergone ring opening (as determined by I H NMR), had a lower tensile strength (11.7 MPa) and resilience (31%), and a higher hardness (51) compared to the corresponding pure material (Table Il). Roux et al. observed a considerable increase in hardness and decrease in tensile strength on going from 10 to 30 mole% epoxidized cis-l,4-polyisoprene, suggesting that in the latter case, ring-opened structures were present.
The nature of the ring-opened products depends on the initial degree of epoxidation. At low modification levels the majority of epoxide groups are isolated, due to the randomness of the reaction and the major ring-opened products are those expected from simple olefin chemistry 6 e.g., hydroxy-acetates, diols and intermolecular ethers, crosslinks. As the modification level (and hence the number of adjacent epoxide groups) increases, the acetate carbonyl absorption
(1728 cm-I ) in the IR spectrum decreases and is replaced by an absorption at 1068
cm-I , previously attributed to a cyclic ether . At 100 mole% modification, the carbonyl absorption was absent.
Intramolecular attack of a hydroxyl group on a neighboring epoxide unit could result in the formation of a 5-, 6-, or 7-membered cyclic ether. The feasibility of this type of reaction has been demonstrated by the attempted epoxidation of 2,6-dimethyloct-2-en-6-ol (I), as shown here:
Even under mild conditions, —IOOC with m-chloroperbenzoic acid in dichloromethane, the epoxide (Il) could not be isolated. Only two products, the five membered cyclic ether (Ill) (76%) and the six membered ether (IV) (24%) were ob-
1 2 3 4 5 6 7
Time/ 103 min
FIG. 7 . —Oxygen absorption of unvulcanized NR and ENR•50 at 1000 C.
tained. Obviously intramolecular cyclization, initiated by a neighboring hydroxyl group, is a very facile reaction. The infrared spectrum of the five-membered product (Ill), 1070 cm-I (C-O) supports this size of cyclic ether in the rubber product, 1068 cm-I . The corresponding absorption in the six membered ether (IV) was observed at 1100 cm
Once ring opening has been initiated, in theory, furan formation should proceed along the rubber backbone until stopped by a nonepoxide group or steric considerations:
This view is substantiated by the titrimetric analysis of ENR with tetramethylammonium bromide. Estimates of epoxide content by this technique were low compared to instrumental methods. However, there was good agreement with the number of epoxide blocks expected from a random epoxidation process25
The product obtained from the reaction of NR latex with 100 mole% peracid is a white amorphous powder, which we have termed "furanized NR". It has a broad Tg around 1000 C, can be injection molded and resembles polystyrene in its impact and stress—strain properties (Figure 8).
CONCLUSIONS AND SUMMARY
These properties indicate that ENR has the potential to become an extremely useful commercial polymer.
The author wishes to thank the Board of MRP RA for permission to publish these results and Mr. M. J. R. Loadman for the analytical data.
6 A. Rosowsky, in "Heterocyclic Compounds with Three and Four Membered Rings, Part I," Inter• science Publishers, New York, 1964.