Liquids that Freeze when Mixed: Homogeneous Cocrystallization Kinetics of Polyoxacyclobutane–Water Hydrate

Abstract

We report on the kinetics of crystallization of polyoxacyclobutane (POCB) and water to form a cocrystal hydrate, the first paper to explore the kinetics of crystallization of a polymer with a small molecule. POCB has unusual cocrystallization behavior when mixed with water, which could be exploited for a variety of applications including improving the nonvolatile memory storage capabilities of carbon nanotube devices, water purification, and biomedical applications such as drug delivery. The rates of hydrate crystallization of a series of mixtures containing up to 24 wt % water were measured using both a bulk, volumetric and a localized, spherulite growth approach. At these compositions, all mixtures were single-phase homogeneous liquids prior to hydrate crystallization, and all mixtures were POCB-rich as compared to the cocrystal stoichiometry. The time dependence of crystallization kinetics is well described by the Avrami equation. Reducing temperature, i.e., increasing undercooling, increases the Avrami exponents and increases spherulite growth velocities consistent with the Hoffmann–Lauritzen model. Reducing water content reduces spherulite growth velocities. The velocities also reduce with time during spherulite growth reminiscent of impurity effects in crystallization. Talc was found to accelerate nucleation. Broadly, the cocrystallization process starting from the homogeneous mixtures of the components resembles homopolymer crystallization, but with the complexity that when the mixture composition deviates from the cocrystal stoichiometry, the excess species cannot crystallize fully.

Salt Effects on the Phase Behavior and Cocrystallization Kinetics of POCB–Water Mixtures

“…The above estimate of the salt fraction of 0.247 is an upper bound, since it assumes full crystallization of POCB, while polymers never crystallize to their full extent. In fact, we noted previously, 6,7 that in the absence of salt, the volume change of crystallization was approximately half of that expected for full hydrate crystallization. Thus, if we take χ = 0.5, eq 1 gives m 0.230 s Lw = , which is still relatively close to full saturation.…”

“…Even a few percent error in the published values of unit cell dimensions change the ρ c value, and hence the value of v c max significantly. 6 Next, we turn to the effect of salt on the crystallization rate. As in our previous studies 6,7 to quantify the kinetics, from each Δv c (t) curve, we identified the time τ at which the volume change reached 0.005 mL/g (chosen arbitrarily), i.e., Δv c (t = τ) = 0.005 mL/g.…”

“…As previously, we seek to quantify the crystallization behavior in terms of two parameters: the value of Δ v c at long times normalΔ v c final , and the time needed to reach a specified value of Δ v c . To estimate normalΔ v c final , we plotted the value of Δ v c at t = 10 4 s (selected arbitrarily) for each experiment.…”

“…A theoretical maximum volume change, normalΔ v c max (assuming all POCB is consumed to form hydrate), can be estimated as normalΔ v normalc max = ( m normalw + m normals ρ aq , normali + m p ρ p ) − ( m POCB 0.764 ρ c + 1 − m POCB 0.764 ρ aq , f ) where ρ aq,i , ρ aq,f , ρ POCB , and ρ c are the densities of the aqueous phase before crystallization, the aqueous phase after crystallization, POCB, and the hydrate crystal respectively, and 0.764 is the weight fraction of POCB in the hydrate crystal. The term in the first brackets is the sum of the specific volume of the saltwater and liquid POCB, which are assumed to be completely immiscible.…”

“…Yet, it should be noted that this conclusion of 50% crystallization depends on the estimated crystal phase density. Even a few percent error in the published values of unit cell dimensions change the ρ c value, and hence the value of normalΔ v normalc max significantly …”