Improved aging and UV resistance of TPEs derived from diimide HSBRs.

In several recent works, the novel preparation of hydrogenated elastomer latexes (HNBR, HSBR and HBR) via an efficient, non-catalytic "diimide" reduction technique has been described (refs. 1 and 2). Furthermore, in the case of hydrogenated styrene-butadiene rubber (HSBR), not only was the rubber latex highly hydrogenated by this technique, the rubber was also found to be a unique thermoplastic elastomer (TPE) with excellent tensile strength as well as outstanding oxidation and ozone resistance (ref. 2).

As with conventional elastomers, however, even hydrogenated TPEs generally require a stabilization system to further improve such critical performance areas as weatherability, use temperature range and processibility (ref. 3). Normally, primary antioxidants, e.g. phenolics or amines, are used often in combination with secondary (synergistic) antioxidants such as phosphites or thioesters to achieve maximum performance. UV protection may also be incorporated by the addition of a UV absorbing agent.

Although it is well known that extensive hydrogenation of diene-type rubbers such as NBR and SBR greatly improves their oxidation and ozone resistance by removing both double bonds and their associated allylic protons, it is less well understood that these hydrogenated polymers can still greatly benefit from proper antioxidant protection. A great deal of information about antioxidant types and their mode of action is available in the literature and is beyond the general scope of this article. However, to the extent that antioxidants can become an integral part of a TPE structure and affect its application performance, their mode of inco oration and benefits will be addressed.

A discussion about polymer-bound antioxidants is again too extensive to be reviewed here, however, several reviews on the subject are available (refs. 4-6). In general, however, the driving force behind polymer-bound antioxidant research is the desire to extend the useful lifetime of a polymer and/or to increase its thermal operating range.

As a rule, conventional antioxidants, especially under aggressive conditions, will be depleted from a polymer by numerous mechanisms such as volatilization, migration, extraction, irreversible absorption on fillers, etc. All these loss mechanisms, however, can be overcome with polymer-bound antioxidants. As a result, polymer properties are better maintained.

Unfortunately, there is little information available in the antioxidant reviews on the use and benefits of bound antioxidants in TPEs, especially TPEs of the SBS or SEBS types. For this reason, it is of interest to determine the relative merits of hydrogenation with and without a bound antioxidant for improving the aged properties of an HSBR TPE.

Our first indication that there might be a performance benefit to bound antioxidants in a hydrogenated elastomer came in 1989 during our initial work on the preparation of HNBR latex (ref 1). During the course of our evaluations, the suitability of various NBR latexes as precursors to HNBR latex was examined. One of these NBR latexes contained the bound amine-type antioxidant, N-(4-anilinophenyl) methacrylamide (APMA) (ref. 7). This latex is used to prepare a commercial NBR rubber. Sequence aging experiments in hot oil and air of aramid-based gasketing sheet made using HNBR latex as a binder showed the latex with the bound antioxidant to be clearly superior (ref. 1).

It remained for this study to determine if the same general performance attributes could be extended to semicrystalline HSBR without adversely affecting their thermoplastic elastomeric properties.

Experimental

Commercially available materials were used in this work except where otherwise specified. Proton-NMR spectra were run on a Varian XL-300 Mhz unit. Infrared spectra were run on a Nicolet fourier transform infrared 20SX-B instrument. Differential scanning calorimetry (DSC) was performed on a Du Pont 9900 or 2200 unit.

Polymerization procedures

The conventional and antioxidant-modified SBRS described in this study were prepared according to the redox recipes shown in table 1. The polymerizations were conducted in quart bottles equipped with punched metal caps and self-sealing gaskets. The bottles were then tumbled in a water bath at 18'C between 7-24 hours to produce latexes from 76-85% conversion. Progress of the individual recipes was monitored by periodically withdrawing samples via syringe and measuring the increase in solids contents with time. The polymerization reactions were then shortstopped at the correct solids by the addition of aqueous solution of 0.1 parts of sodium N, N-dimethyldithiocarbanate and 0.1 parts of N,N-dimethylhydroxylamine. The shortstopped latexes werer then vacuum stripped to remove residual monomers. The polymerization and raw polymer data for these experimental SBSs area show in table 2

[TABULAR DATA OMMITTED]

Table 2 - polymerization and polmer data for SBRs Polymer A B C D E F

Polymer Class SBR SBR SBR SBR/ SBR/ SBR/

AO AO AO Pzn.time, h. 7.0 7.0 7.0 24.0 24.0 24.0 Conversion, % 76 81 80 85 85 82 Mooney Viscosity, 15 19 12 77 69 63 ML-4 @ 100C

Composition by[1] H-NMR, wgt%: Butadiene Styrene  95.2 91.1 87.9 94.7 92.3 87.3 (mole%) 4.8 8.9 12.1 4.8 7.2 12.7 APMA* (2.5) (4.8) (6.7) (2.5) (3.9) (6.7)

-- -- -- 0.5 0.5 0.5

*N-(4-anilinophenyl) methacrylamide

Latex reduction process (LRP) The stripped latexes were then used directly in the diimide reduction process to produce die desired hydrogenated SBRS both with and without the polymer-bound antioxidant. The general experimental procedure and advantages of this reduction technique over conventional polymer hydrogenation methods have been previously published (refs. 1 and 2). The diimide-reduced polymers were subsequently analyzed by FRIR and DSC methods and the results are summarized in table 3.

[TABULAR DATA OMMITTED]

Results and discussion

Mechanism of the latex reduction process Our most recent insight into the mechanism of the micellar reduction technique is focused upon answering the fundamental question of how the short-lived and very reactive diiniide molecule is so efficiently generated and consumed in double bond reduction. The process appears to be almost enzyme-like in behavior - and this behavior provides the necessary clue.

Extensive published research on enzyme active site structures such as those found in hemerythrin, ribonucleotide reductase, methane monooxygenase and many other enzymes reveals the presence of binuclear metallic cores, many of which are carboxylate-bridged (refs. 8-13). Additionally, it is also known that metal-containing liquid crystals in the form of metal binuclear carboxylates can be readily formed from the simple interaction of long chain sodium carboxylates and simple transition metal salts such as copper sulfate (ref 14).

Since it is known that an anionic so (sodium or sium fatty acid carboxylate) and copper ion are both present and required to give an efficient diimide latex reduction reaction (ref 1), it is most probable that the active metal site at the latex particle surface also has the same binuclear structure. If this is true, then the binuclear copper soap complex resides at the latex particle surface where it "sees" a large excess of hydrazine in the bulk aqueous phase of the latex. It is very likely that hydrazine binds between the copper centers of the binuclear complex through the lone electron pairs of the hydrazine nitrogen atoms. Oxidation of the bound hydrazine with hydrogen peroxide then immediately produces diimide which itself may be stabilized to some extent by interaction with the dicopper site.

There are numerous examples in the literature where the diimide molecule is stabilized in this bimetallic fashion (refs. 15 and 16), although no stable dicopper diimide complexes have been reported (ref

Although the circumstantial evidence strongly indicates that diimide is generated at a binuclear copper site, it is still unclear exactly how the polymeric olefin groups interact to become reduced. One possibility is that excess hydrazine in the reaction medium could act to displace the coordinated diimide due to mass action, and/or a higher complexation constant. In this way, uncomplexed diimide could be generated. Reaction of the olefin with uncomplexed diimide would not then be subject to an entropically unfavorable polymer/ dicopper-diimide complex interaction geometry.

SBRIHSBR preparation and characterization The polymerization recipes used to prepare the SBR polymers for this study are shown in table I and the pertinent polymerization and raw polymer data are shown in table 2. From the data in table 2, several interesting observations can be made. Although all the polymer conversions are similar, the presence of the antioxidant monomer greatly increased the polymerization time (from 7-24 hours) and increased the raw polymer Mooney viscosity (from about 15 to 70). The reason for the decreased rate of polymerization is not entirely clear, but it is known that this particular antioxidant monomer interferes with the iron of the redox initiator system (ref 7). This effect usually can be compensated for by increased iron levels in the recipe (table 1), but was not entirely successful in this case. In addition, the higher Mooney viscosity of the antioxidant-containing polymers may be caused by interchain hydrogen bonding between the pendent amine-amide functional groups.

The experimental hydrogenated polymers derived from the SBRS (A through F) of table 2 were prepared and evaluated to answer two fundamental questions: How does a polymer bound antioxidant affect the physical properties and aging behavior of diimide-hydrogenated SBR rubber versus hydrogenated SBR without bound antioxidant? and, How does varying the mole % of styrene in the HSBR (with or without bound antioxidant) affect the physical properties of the polymers at intermediate levels of hydrogenation (83-89%)?

Using the latex reduction technique, the SBR latexes were converted to HSBR latexes (AR thru FR). The hydrogenation levels of these polymers were determined by an FTIR method and are shown in table 3. For comparison purposes, it should be noted that sample pairs AR/DR, BR/ER and CR/FR each contain approximately the same mole % of styrene, but differ in antioxidant content.

The DSC (nitrogen) data from table 3 indicate little difference in Tg between any of these polymers; ranging between -39.7 and 45.6[degrees centigrade]. This indicates there is little structural difference in the amorphous portion of these polymers. In contrast to this, however, the energy of the crystalline melting endotherm shows (as expected) the AR/DR pair to have the most crystallinity and the CR/FR pair to have the least. This is because the AR/DR pair has the lowest styrene content and therefore the highest level of polyethylene-like units. Although the endotherm energy for DR is somewhat less than that for AR (34.5 vs. 40.0 J/g), this difference can be most easily explained by the difference in hydrogenation levels (86.4 vs. 88.7) rather than an effect of the antioxidant. This effect can also be seen in the CR/FR pair where FR has the lower degree of hydrogenation and the smaller melting endotherm.

Oxidation Resistance The similarity between these groups of polymers disappears, however, when they are extracted and evaluated by thermooxidative DSC. The thermooxidative data shown in the lower part of table 3 clearly shows a major difference between samples AR and CR without bound antioxidant on the one hand (exotherm onsets between 142-158[degrees centrigrade] and samples DR and FR with a bound antioxidant (exotherm onsets of 218[degrees centigrade] on the other. Additionally, variations in styrene content and levels of hydrogenation within the range studied show little or no effect on the oxidative onset temperature. The DSC scans of HSBRs containing the bound antioxidant (0.5 phr bound) all show a similar pattern with a single sharp and very large exotherm (figure 1). We interpret this result as being indicative of an excellent bound antioxidant distribution in the amorphous polymer phase resulting in essentially only one uniformly oxidizable environment. Note: the antioxidant molecule (and styrene) are too large to be incorporated into polyethylene crystallites.

In contrast to the DSC thermogram shown in figure 1, all the HSBRS without the bound antioxidant also show a typical pattern (figure 2) with a much lower oxidative onset temperature and multiple broad exotherms. We interpret this behavior as being indicative of a complex autocatalytic oxidation chemistry.

It is fortuitous in the case of the HSBRS with bound antioxidant, that the residual unsaturation and bound antioxidant are both concentrated in the same amorphous region of the polymer since this region is the most prone to oxidation.

To verify the excellent oxidation resistance predicted by the thermooxidative DSC data, latex cast films of polymers AR, CR, DR and FR were prepared and evaluated for both their initial and aged physical properties. The results of this evaluation are shown in table 4. Once again, little difference is seen among their initial properties except for sample FR which has a somewhat lower tensile strength and higher elongation than the others. As mentioned previously, this difference is most likely due to its slightly lower saturation level (83%) and not due to the presence of the bound antioxidant.

Forced-air oven aging of these cast latex films at 1400C for 65 hours, however, produced a dramatically different picture. Here, the tensile strength and elongation of samples AR and CR without the bound antioxidant deteriorated badly (7% retained elongation), whereas the properties of DR and FR with antioxidant were still quite respectable at 75% retained elongation. Table 4 - physical property data for intial and

aged HSBR latex cast films HSBR A-R C-R D-R* F-R* % Butadiene hydrogenation: 88.7 85.5 86.4 83.0

Initial properties: Tensile, MPa 11.5 11.1 11.5 8.4 Elongation, % 744 776 729 833 Modulus, MPa 200% 3.0 2.7 3.0 2.0 300% 3.3 3.1 3.4 2.3 500% 4.5 3.8 4.6 2.9

Aged prpoperties, 65 h @ 140[degrees centigrade in forces air oven : Tensile, MPA 3.8 2.7 8.3 6.2 Elongation, % 53 55 535 622 % Elongation retained 7 7 73 75 Modulus, MPa 200% -- -- 3.2 2.0 300% -- -- 3.9 2.5 500% -- -- 7.1 3.9

*Contain the polymer-bound antioxidant (APMA)

UV Resistance Samples of the latex cast films used for the oven aging tests were also evaluated for their resistance to ultraviolet degradation in an accelerated weatherometer test. The conditions of this test and the results are shown in table 5. Once again, as with the air oven aging results, samples DR and FR showed much superior tensile and elongation retention versus samples AR and CR. We found these results to be somewhat surprising for the antioxidant-containing HSBRS, even though the antioxidant monomer is reported to have a [lambda] max of 310 nm and an extinction coefficient (E) of 2.3 X 104 (ref. 7). Often, many common classes of antioxidants that function well under thermooxidative conditions show poor performance under photooxidative ones ref. 18).

Summary

Several interesting features to the thermooxidative and photooxidative behavior of HSBRS with and without a polymer bound antioxidant have been observed. Most important is the observation that HSBRS containing a bound amine-type antioxidant and moderate hydrogenation levels (83-86%) can greatly outperform HSBRS with similar saturation levels without the bound antioxidant with regard to aged property retention under thermal and photolytic conditions. It also appears that the HSBRS with a bound antioxidaot and these modest hydrogenation levels can approach the thermal resistance of even very highly saturated (97%) HSBR (without antioxidant) (ref. 2).

Secondarily, die incorporation of a polymer bound antioxidant into an emulsion SBR and subsequently into an HSBR does not interfere with the latex diimide reduction process or the ultimate generation of TPE properties. The antioxidant molecule, because of its large size, must be excluded from the polyethylene-like crystallites and concentrate in the amorphous rubbery domains where antioxidant protection is most needed.

Additionally, variation of styrene levels between 2.5 and 6.7 mole % in the HSBRS studied had little influence on polymer aging properties, but was noticeable in affecting the initial level of polymer crystallinity as measured by DSC.

In conclusion, we believe our observations concerning the benefits of bound antioxidants in HSBR hold important implications with regard to the quality and recyclability of TPES in general. It is clear from this work that antioxidant protection is required even for highly saturated, semicrystalline polymers. Since most TPES are utilized at 50% or less by weight in a fully compounded stock, the likelihood that conventional antioxidants will remain exclusively with the polymer phase is remote. In turn, loss of an effective amount