No abstract

Investigating the compression properties of randomly ordered fiber networks experimentally is difficult which has resulted in ongoing disputes as to the mechanisms controlling the compression strength in such materials. In this work, we investigated compression properties of randomly oriented fiber networks with a special emphasis on cellulose products such as paperboard. We numerically reconstructed the conditions of the short span compression test widely used to quantify the compression strength of paperboard. We found that the phenomenological failure mode of such networks is elasto-plastic buckling. The x-shaped failure mode observed in physical experiments appears when test specimen restraints are included in the model. The most significant improvements to sheet strength can be obtained by improving the elastic properties while the strain to failure is increased most by an improvement of the plastic yield and hardening properties of individual fibers. Bond breaks were confirmed to have a smaller influence on the overall response. Fiber level microscopic buckling was investigated in depth, providing quantitative estimates of the fraction of mass likely to buckle at the microscopic level. The analysis indicated that only a low to moderate number of load carrying fibers can be expected to buckle. The inherent strength reserve in non-ordered fiber networks was investigated by introducing hinge mechanisms throughout the network, and the effect was shown to be small for a small to moderate number of hinges.

Cellulose fibers are a major industrial input, but due to their irregular shape and anisotropic material response, accurate material characterization is difficult. Single fiber tensile testing is the most popular way to estimate the material properties of individual fibers. However, such tests can only be performed along the axis of the fiber and are associated with problems of enforcing restraints. Alternative indirect approaches, such as micro-mechanical modeling, can help but yield results that are not fully decoupled from the model assumptions. Here, we compare these methods with nanoindentation as a method to extract elastic material constants of the individual fibers. We show that both the longitudinal and the transverse elastic modulus can be determined, additionally enabling the measurement of fiber properties in-situ inside a sheet of paper such that the entire industrial process history is captured. The obtained longitudinal modulus is comparable to traditional methods for larger indents but with a strongly increased scatter as the size of the indentation is decreased further.