6970
Cellulose (2020) 27:6961–6976
bonding but stiff and long hemp fibres. The compres- sion stress and modulus of the PVA trial point were further improved with the addition of hemp (TP6). In contrast, a slight drop in these properties was seen for the SDS trial point (TP3) where fibres distributed more evenly and the local densification was not as strong, see Fig. 7a. When looking at the stress–strain curves of Fig. 8, the refined and unrefined NBSK samples obtained with SDS follow the predicted theoretical behaviour (Ketoja et al. 2019) of random fibre networks [refer to Eqs. (1, 2)] quite well, except below 10% strain (Ma¨kinen et al. 2020). A slight deviation was observed with the added hemp component in Fig. 8c, which could have resulted from either voids in the structure or altered fibre segment distribution for intermediate span lengths. Interestingly, similar compression-stress behaviour was produced by unrefined NBSK with the PVA foaming agent (see Fig. 8d), which produced an uneven fibre network with large voids (Fig. 7b). It is possible that PVA forms films (Bossu et al. 2019) that, together with increased bonding and locally increased density, interfere with the distribution of intermediate segment lengths in a fashion similar to when refining and hemp are added. The deviation from the model behaviour for the PVA samples appeared to increase with NBSK refining (Fig. 8e) and added hemp com- ponent (Fig. 8f). Both factors probably drove the formed complex multi-scale material network away from the simple random fibre network that had exponentially distributed free segment lengths. The scaled slope of stress–strain curve 1 r d r d at 50% com- pression, when the large voids can be expected to be closed, varies in the range of 2.8–4.4, see Table 2. All
results with SDS surfactant are relatively close (average deviation 9%) to the theoretical prediction 3.8 given by Eq. (5). The larger deviations from the theory with PVA foaming agent for refined NBSK cause a clear drop in the slope at 50% compression. Buckling of fibre segments has been postulated as the failure mechanism underlying Eqs. (1, 2) (Ketoja et al. 2019; Ma¨kinen et al. 2020). Thus, the recovery of sample thickness after the load has been removed should depend mainly on the recovery of fibres from buckling deformation and not on bonding properties. Figure 9 shows the thickness recovery of samples exposed to 50% compression recorded 1 min after the load was removed. The recovery was roughly equal for both surfactant types despite their highly different bonding properties. This agrees with the above theoretical postulate. Unrefined curly Kraft fibres led to better recovery than straighter refined Kraft fibres. Addition of hemp fibres (TP3 and TP6) did not change this picture.
Effects of lignocellulosic fines, cellulose nanofibrils, and LBG
The results presented in the previous chapter suggest that the fines and fibrillation coming from refining, together with added strengthening polymers like PVA, can ‘‘activate’’ the multi-scale fibre network and improve strength properties significantly. Thus, the contribution to strength of fines, fibrils, and macro- molecules was studied more thoroughly. This was done for the fibre composition that included 80% NBSK and 20% long hemp pulp fibres. The testing plan (see Table 3) included the reference with starch
Fig. 6 a Specific compression stress (compression stress divided by bulk density) of the materials (Table 2) made with SDS (TP1–3) and PVA (TP4–6) foams at 10% and 50% deformation. b The compression modulus of the same materials
123
Made with FlippingBook Online newsletter maker