Phase Behavior of "Block-Random" Copolymers

            “Block-random” copolymers—represented generally by the structure (A_xB_1-x)-(A_yB_1-y), where each of the two blocks is a random copolymer of monomers A and B, simply with different fractions of A (x and y)—present a convenient and useful variation on the typical block copolymer architecture, as the interblock interactions and physical properties can be tuned continuously, via x and y, through the random block's composition. The ability to tune the effective interaction parameter between the blocks continuously, allows for the order-disorder transition temperature (T_ODT) to be tuned independently of molecular weight using only two monomers, via the difference¹ between x and y. This flexibility makes block-random copolymers a versatile platform for the exploration of polymer phase behavior and structure-property relationships.

Typical living or controlled polymerizations produce compositional gradients along the random block, since the reactivity ratios are generally different from unity, which can in turn influence the phase behavior. Living polymerizations which proceed with random monomer addition (no gradient) are consequently sought because they would permit the synthesis of well-defined polymers, and block copolymers, of tunable composition (hence tunable properties) which are effectively homogeneous on length scales larger than the monomer size.² It is desirable to achieve such random copolymerizations of styrene and isoprene (SI), to expand the accessible range of polymer properties. In particular, hydrogenated high-vinyl polyisoprene has an exceptionally low cohesive energy density,³ lower than that for hydrogenated high-vinyl polybutadiene, meaning that a broad range of solubility parameters (>2 MPa^1/2) could be accessed in random copolymers containing styrene and hydrogenated vinyl isoprene (hI) units. This increased range of solubility parameters translates directly into an increased range of accessible values of the interblock Flory-Huggins interaction parameter χ in block-random copolymers.

The reactivity ratios for SI in hydrocarbon and ether solvents are qualitatively similar to those for SB,⁴ but despite the seemingly minor difference between butadiene and isoprene, N,N,N¢,N¢-tetramethylethylenediamine, TMEDA is not an effective randomizer for SI copolymerizations.⁵ Over half a century ago, Kelley and Tobolsky⁶ showed that triethylamine (TEA) is an excellent randomizer for equimolar SI copolymerizations (~60 wt% S). Later, Annighöfer and Gronski⁷ used a TEA/benzene mixture (20/80 v/v) to synthesize S-SrI-I triblock copolymers with styrene-ran-isoprene (SrI) midblocks (~50 wt% S); they reported reactivity ratios r_I = 1.0 and r_S = 0.8, implying that effectively random copolymers could be synthesized at essentially any S:I ratio with this approach.

Here, organolithium initiation in a cyclohexane/triethylamine mixture yields narrow-distribution copolymers of styrene and isoprene of any desired composition, with no measurable down-chain gradient. These random copolymers (SrI) have been successfully incorporated into well-defined symmetric block copolymers (I-SrI diblocks) and subsequent isoprene-selective hydrogenation yields thermally stable hI-SrhI diblocks, which self-assemble into well-defined lamellar morphologies with sharply-defined order-disorder transitions, whose temperatures scale predictably with diblock molecular weight. The use of SrhI in lieu of a styrene homopolymer block effectively dilutes the unfavorable contacts between the two blocks in the homogeneous phase reducing the effective X, where X if the interaction energy density (Χ = χ(Nρ/M)RT = (χN)_ODTρRT_ODT/M) by a factor of ~8, and allowing for greatly elevated molecular weights and d spacings at a given value of T_ODT. The measured interaction energy density between hI and SrhI is consistent with the mean-field “copolymer equation”⁸, predicts factor of ~5 reduction in X, providing a first step towards the design of styrene-isoprene block-random copolymers of desired molecular weight and T_ODT.

Within the context of mean-field theory, as in the Flory-Huggins model, X is related to the difference in solubility parameters (d) of the two blocks as X=(Δd)². Having measured values of X we can rank SrhI in solubility parameter relative to other polymers, particularly the saturated hydrocarbon polymers studied extensively by Graessley and coworkers.^3,9 We find that our determined solubility parameter for SrhI matches well with what would be calculated from homopolymer solubility parameters, for hI and S, using a volume fraction weighted average. Interestingly, the value of d_SrhI determined here is close to that for polyethylene,⁹ suggesting potential miscibility between polyethylene and SrI, an idea we are in the process of testing.

This work was generously supported by the National Science Foundation, Polymers Program (DMR-1003942).

References

(1)        Smith, S. D.; Ashraf, A.; Satkowski, M. M.; Spontak, R. J. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1994, 35(1), 651-652.

(2)        Tierney, N.K.; Register, R. A. Macromolecules 2003, 36, 1170-1177.

(3)        Graessley, W.W.; Krishnamoorti, R.; Reichart, G. C.; Balsara, N. P.; Fetters, L. J.; Lohse, D. J. Macromolecules 1995, 28, 1260-1270.

(4)        Morton, M. Anionic Polymerization: Principles and Practice; Academic Press: New York, 1983, pp. 134-145.

(5)        Chang, C. C.; Halasa, A. F.; Miller, Jr., J. W.; Hsu, W. L. Polym. Int. 1994, 33, 151-159.

(6)        Kelley, D. J.; Tobolsky, A. V. J. Am. Chem. Soc. 1959, 81, 1597-1600.

(7)        Annighöfer, F.; Gronski, W. Coll. Polym. Sci. 1983, 261, 15-25.

(8)        Graessley, W. W. Polymeric Liquids & Networks:  Structure and Properties; Garland Science:  New York, 2004, pp. 341-407.

(9)        Krishnamoorti, R.; Graessley, W. W.; Dee, G. T.; Walsh, D. J.; Fetters, L. J.; Lohse, D. J. Macromolecules 1996, 29, 367-376.