Acetylation of Nanocellulose: Miscibility and Reinforcement Mechanisms in Polymer Nanocomposites

The improvement of properties in nanocomposites obtained by topochemical surface modification, e.g., acetylation, of the nanoparticles is often ascribed to improved compatibility between the nanoparticle and the matrix. It is not always clear however what is intended: specific interactions at the interface leading to increased adhesion or the miscibility between the nanoparticle and the polymer. In this work, it is demonstrated that acetylation of cellulose nanocrystals greatly improves mechanical properties of their nanocomposites with polycaprolactone. In addition, molecular dynamics simulations with a combination of potential of mean force calculations and computational alchemy are employed to analyze the surface energies between the two components. The work of adhesion between the two phases decreases with acetylation. It is discussed how acetylation can still contribute to the miscibility, which leads to a stricter use of the concept of compatibility. The integrated experimental-modeling toolbox used has wide applicability for assessing changes in the miscibility of polymer nanocomposites.

transformation of a single surface OH into OAc, which gives the total free energy change ∆ of that process.The absolute value calculated in this way has no direct physical interpretation as there is no reference state.However, from two separate simulations performed for surfaces in contact with both polymer and vacuum, respectively, one can express the change in adhesion from acetylation as (Figure S4B) where A is the surface area.
It was evaluated for two separate cases: one where only one of the C6 groups were acetylated, and one where they were all acetylated.For reference, Δ 12 (0) was also calculated for a surface in contact with water, at the same temperature (the same used for the melt processing, 393 K).The changes in (absolute) free energy from the computational alchemy simulations are given in Table S5.Here, to compute a total change in CNC-PCL adhesion from acetylation, we use Δ 12 (0) = 〈∆∆〉, (2) where  is the surface number density of available C6 groups, and the average ∆∆ per surface group is taken as the mean of the calculated values at the low and the high degrees of acetylation.The average ∆∆ is -7.3 kJ mol -1 and -4.3 kJ mol -1 for CNC-PCL and CNC-water, respectively.The crystal lattice parameters of native cellulose give  = 1.8 nm -2 for the [1-10] plane.  .X-ray diffractograms of CNC (red curve) and acetylated AcCNC (blue curve).Characteristic peaks presenting crystalline structure of cellulose I appear at 2: 14.7, 16.8, and 22.7, and a very small peak ascribable to cellulose II emerge at 2: 21.9.From a qualitative comparison of the diffractograms both cellulose nanocrystals present mainly in crystalline structure of cellulose I, and no significant changes can be observed in the CNC crystalline structure after one-pot acid hydrolysis-Fisher esterification used for the topochemical acetylation at the CNC surface.

Figure S4
Illustration of the computational alchemy approach.(A) One surface hydroxyl group is "mutated" into an acetyl group by the use of dummy atoms, and the free energy of that process is calculated from the simulations.(B) The surface acetylation is simulated for three different cases: when the crystal surface is in contact with a PCL melt, vacuum, or liquid water, respectively.The cellulose slab is viewed along the chain axis.The difference in work of adhesion (ΔWA) between acetylated/non-acetylated systems is related to the calculated free energies since any closed cycle must add up to zero (see Eq. S2).(C) The low-acetylated (top) and high-acetylated (bottom) surface models used in this work.Tables Table S1 -Representative degradation temperature Td and onset of the degradation at the 5% of weight loss, T5wt%, from TGA measurements of pristine CNC, AcCNC, and melt processed PCL matrix, and nanocomposites.Table S3.Tensile properties of the melt processed PCL matrix, and nanocomposites (Young's modulus, ultimate strength, strain to failure and work to fracture).

Sample E young [MPa]


Figure S2 .
Figure S2.DSC curves for the nanocomposites and the neat PCL (black curve), 10%CNC/PCL (green curve) and 10%AcCNC/PCL (lime curve), run from -80°C to 140°C corresponding to the second heating used for the herein analysis.Integration region chosen for all sample for the assessment of enthalpies and calculation of degree of crystallinity was as shown in figure (pink dashed rectangular shape) between 6°C and 70°C i.e. over the same temperature interval.Following data were achieved from calculations for the degree of crystallinity using the equation  = (ΔH m /(xPCL* ΔH m 0 )*100, where ΔH m is the melting enthalpy measured in the DSC second heating scan, x is the weight fraction of PCL in the nanocomposites, and enthalpy value for 100% crystalline PCL ΔH m

Figure S5
Figure S5Density profiles of PCL at CNC (green), AcCNC (dashed green) interfaces compared to a PCL interface to air (blue).

Figure S6
Figure S6Histograms of the center-of-mass separation from the umbrella sampling simulations.There is sufficient overlap to ensure convergence of the resulting PMF.

Table S2 -
Value assessed from the minimum in the 1 st derivative curves Determined from TGA under N2 with a heating rate of 20°C/min.DSC main results, degree of crystallinity, melting temperature Tm and enthalpy ΔHm.