**4. Cellulose–epoxy composites**

**3.3. Immobilization of lignosulfonate**

206 Composites from Renewable and Sustainable Materials

reduce the heat released during combustion.

different batches of ethoxylated lignosulfonate.

To prevent the migration of the lignosulfonate, it was epoxidated using epichlorohydrin. It was then used to prepare epoxy + 10 % eLS composites, which were still opaque, but lighter in color than the REAX825E composites. The addition of glycidyl moieties did add to the fuel content of the lignin. Using MCC, eLS was found to have Hc = 5.1 kJ/g, HRC = 72 J/g · K, Tpeak = 340 °C, and 44 % mass fraction residue. This is only a modest increase in flammability and the immobilization was expected to have a greater effect. As shown in **Figure 3**, by incorporating the lignosulfonate within the epoxy matrix, the char layer was more effective at reducing the heat released. The PHRR was reduced by 63 % and the THR was reduced by 27 % over neat epoxy. The one disadvantage is that the composite ignited the earliest of all composites tested. To improve the efficiency of the epoxidation process, the starting material was dissolved in ethanol and filtered to remove any insoluble material. The solubility of lignin in ethanol decreases as the molecular weight of lignin increases, so this likely removes the highest molecular weight lignin. This may account for the earlier time to ignition, though residual ethanol will also lower this characteristic. The char formed during the decomposition was thicker and did not peel away from the composite, which helps explain how it could effectively

**Figure 2.** Heat release rate as measured in a cone calorimeter of epoxy flame retarded with lignosulfonate, using two

The addition of lignosulfonate significantly changed the glass transition temperatures (Tg). The measured Tg were 81.2 °C for neat epoxy, 88.7 °C for epoxy + 10 % REAX825E, and 75.7 °C for epoxy + 10 % eLS. Increases in Tg are typically associated with enthalpic interactions between

#### **4.1. Effects of processing methods of dispersion**

The order of addition and the use of high speed centrifugal mixing were investigated to improve dispersion of unmodified sulfated cellulose nanocrystals (Na-CNC) in epoxy composites. The Na-CNC nanocrystals that were obtained as a dried powder were freeze dried in the presence of 9 % mass fraction t-butyl alcohol. This reduces the water crystal size during the freezing process and adds a small amount of an organic solvent to the crystal structure of the cellulose, which improves the microscopic dispersion of the crystals. Since these represent the "best case scenario" for dispersing unmodified sulfated cellulose nanocrystals in hydrophobic matrices, these were used to assess the effects of processing methods on dispersion. All the composites exhibited cellulose aggregation that was visible without magnification. Mechanical stirring and sonication resulted in the largest aggregates and high shear mixing with addition of cellulose to the diamine followed by addition of the epoxy resin resulted in the smallest aggregates. In previous studies, both a polar, nonaqueous solvent and ultrasonication were required to form transparent composites free from visible aggregates.

#### **4.2. Ion exchange of cellulose nanocrystals**

A method has recently been developed to modify the surface of sulfated cellulose nanocrystals using a simply, scalable ion exchange process [41]. In this approach, rather than adding a surfactant, the sodium cation is exchanged with a surfactant cation. The exchange reduces the surface energy, reduces the water uptake, increases the thermal stability, and improves the polymer adhesion of the nanocrystals. This method also allows co-exchange of cations. In this study, the column was loaded with 1 % mass fraction rhodamine 6G and 99 % mass fraction methyl(triphenyl)phosphonium, (MePh3P/Rh)-CNC. A Na-CNC control was prepared by using a column that was loaded with 1 % mass fraction rhodamine 6G and 99 % mass fraction Na+ . The fluorescence of 2 % mass fraction cellulose nanocrystals in water is shown in **Figure 4**.

**Figure 4.** UV and fluorescence spectra of aqueous solutions of cellulose nanocrystals.

#### **4.3. Improved dispersion with exchanged cellulose nanocrystals**

The exchanged cellulose nanocrystals disperse readily in the epoxy resin, even without significant shear. Composites were initially prepared using a high shear, speed mixer. Optical images (**Figure 5**) show large microscopic agglomeration for the Na-CNC composites, which disappear when using MePh3P-CNC. There is still some agglomeration, but the lengths of the aggregates are smaller than 50 μm and the width of the aggregates is on the submicron scale.

the smallest aggregates. In previous studies, both a polar, nonaqueous solvent and ultrasoni-

A method has recently been developed to modify the surface of sulfated cellulose nanocrystals using a simply, scalable ion exchange process [41]. In this approach, rather than adding a surfactant, the sodium cation is exchanged with a surfactant cation. The exchange reduces the surface energy, reduces the water uptake, increases the thermal stability, and improves the polymer adhesion of the nanocrystals. This method also allows co-exchange of cations. In this study, the column was loaded with 1 % mass fraction rhodamine 6G and 99 % mass fraction methyl(triphenyl)phosphonium, (MePh3P/Rh)-CNC. A Na-CNC control was prepared by using a column that was loaded with 1 % mass fraction rhodamine 6G and 99 % mass frac-

. The fluorescence of 2 % mass fraction cellulose nanocrystals in water is shown in

cation were required to form transparent composites free from visible aggregates.

**Figure 4.** UV and fluorescence spectra of aqueous solutions of cellulose nanocrystals.

**4.3. Improved dispersion with exchanged cellulose nanocrystals**

The exchanged cellulose nanocrystals disperse readily in the epoxy resin, even without significant shear. Composites were initially prepared using a high shear, speed mixer. Optical

**4.2. Ion exchange of cellulose nanocrystals**

208 Composites from Renewable and Sustainable Materials

tion Na+

**Figure 4**.

**Figure 5.** Optical images (20× magnification) of (a) 0.5 % Na-CNC in epoxy and (b) 0.5 % MePh3P-CNC in epoxy.

When using an optical microscope, it is often difficult to differentiate cellulose crystals from matrix defects, such as cracks, bubbles, or impurities. To obtain a better representation of cellulose distribution, laser scanning fluorescent confocal microscopy was used to image epoxies containing rhodamine co-exchanged crystals (cf **Figure 6**). These images show larger Na-CNC agglomerates, since the (Na/Rh)-CNCs were freeze dried without t-butyl alcohol. They also show fewer microscopic (MePh3P/Rh)-CNC than the optical images of MePh3P-CNC containing epoxies (**Figure 5b**). The visible crystal sizes are similar to those observed for MePh3P-CNC. This suggests that some of the features observed in the light microscope images are defects rather than cellulose crystals. It also indicates that there are crystals that are smaller than the limit of detection of optical microscopes, likely nanoscale in size. This mixed microscale/nanoscale size distribution of nanocellulose in polymer matrices has been observed previously [42].

**Figure 6.** Fluorescent confocal images (100× magnification) of (a) 0.5 % Na-CNC in epoxy and (b) 0.5 % MePh3P-CNC in epoxy.

Composites were also prepared by gently heating the epoxy resin to reduce viscosity, blending the cellulose using a stir plate, and use of a sonication bath to help separate some of the slower dispersing particles. Epoxies containing 2 % MePh3P-CNC were visibly transparent, and the microscopic images (not shown) suggest similar levels of crystal separation and dispersion throughout the epoxy.

The differences in particle size and surface energy of the crystals resulted in differences in the glass transition temperatures (cf **Figure 7**). The cellulose nanocrystals interacts strongly with the epoxy matrix, leading to a 5 °C increase in the glass transition temperature at 2 % NaCNC loading. This behavior is consistent with other studies incorporating cellulose nanofibers [43] or nanocrystals [44]. The addition of MePh3PCNC does not change the glass transition temperature relative to neat epoxy. The MePh3P+ functionality lowers the surface energy of the cellulose, reducing both crystal–crystal interactions (less aggregation), and cellulose– polymer interactions (lower Tg). The differences in Tg between the epoxies prepared using high shear mixing and those using mechanical mixing were not statistically significant.

**Figure 7.** Glass transition temperatures of epoxy-cellulose nanocrystal composites using DSC. Error bars represent 2σ.

The tensile properties of the epoxy composites are provided in **Table 2**. As expected, the incorporation of stiff crystals increased the modulus (E) of the epoxy composites. The addition of MePh3P-CNC had a greater effect than the addition of Na-CNC, likely due to the smaller size of aggregates and better dispersion in these composites. The peak tensile strength (σp) of the composites decreased, which may be due to heterogeneity in the cross-linking due to the presence of the crystals. The use of MePh3P-CNC also reduced the tensile strength, but only half as much in the Na-CNC composites. A key advantage to using the modification is the reduction in hydrophobicity due to the presence of large, hydrophobic cations. The water absorption in Na-CNC epoxy after 1 day of water immersion (W1d) was significantly higher than in neat epoxy, but the MePh3P-CNC filled epoxies had nearly the same water absorption as neat epoxy.


**Table 2.** Tensile properties and water absorption of cellulose nanocrystal–epoxy composites.
