The need for multifunctional, robust, reusable, and high-flux filters is a constant challenge for sustainable water treatment. In this work, fully biobased and biodegradable water purification filters were developed and processed by the means of three-dimensional (3D) printing, more specifically by fused deposition modelling (FDM). The polylactic acid (PLA) – based composites reinforced with homogenously dispersed TEMPO-oxidized cellulose nanofibers (TCNF) or chitin nanofibers (ChNF) were prepared within a four-step process; i. melt blending, ii. thermally induced phase separation (TIPS) pelletization method, iii. freeze drying and iv. single-screw extrusion to 3D printing filaments. The monolithic, biocomposite filters were 3D printed in cylindrical as well as hourglass geometries with varying, multiscale pore architectures. The filters were designed to control the contact time between filter's active surfaces and contaminants, tailoring their permeance. All printed filters exhibited high print quality and high water throughput as well as enhanced mechanical properties, compared to pristine PLA filters. The improved toughness values of the biocomposite filters clearly indicate the reinforcing effect of the homogenously dispersed nanofibers (NFs). The homogenous dispersion is attributed to the TIPS method. The NFs effect is also reflected in the adsorption capacity of the filters towards copper ions, which was shown to be as high as 234 and 208 mg/gNF for TCNF and ChNF reinforced filters, respectively, compared to just 4 mg/g for the pure PLA filters. Moreover, the biocomposite-based filters showed higher potential for removal of microplastics from laundry effluent water when compared to pure PLA filters with maximum separation efficiency of 54 % and 35 % for TCNF/PLA and ChNF/PLA filters, respectively compared to 26 % for pure PLA filters, all that while maintaining their high permeance. The combination of environmentally friendly materials with a cost and time-effective technology such as FDM allows the development of customized water filtration systems, which can be easily adapted in the areas most affected by the inaccessibility of clean water.

Cellulose fibrils are the structural backbone of plants and, if carefully liberated from biomass, a promising building block for a bio-based society. The mechanism of the mechanical release-fibrillation-is not yet understood, which hinders efficient production with the required reliable quality. One promising process for fine fibrillation and total fibrillation of cellulose is cavitation. In this study, we investigate the cavitation treatment of dissolving, enzymatically pretreated, and derivatized (TEMPO oxidized and carboxymethylated) cellulose fiber pulp by hydrodynamic and acoustic (i.e., sonication) cavitation. The derivatized fibers exhibited significant damage from the cavitation treatment, and sonication efficiently fibrillated the fibers into nanocellulose with an elementary fibril thickness. The breakage of cellulose fibers and fibrils depends on the number of cavitation treatment events. In assessing the damage to the fiber, we presume that microstreaming in the vicinity of imploding cavities breaks the fiber into fibrils, most likely by bending. A simple model showed the correlation between the fibrillation of the carboxymethylated cellulose (CMCe) fibers, the sonication power and time, and the relative size of the active zone below the sonication horn.

Nanocelluloses are seen as the basis of high-performance materials from renewable sources, enabling a bio-based sustainable future. Unsurprisingly, research has initially been focused on the design of new material concepts and less on new and adapted fabrication processes that would allow large-scale industrial production and widespread societal impact. In fact, even the processing routes for making nanocelluloses and the understanding on how the mechanical action fibrillates plant raw materials, albeit chemically or enzymatically pre-treated, are only rudimentary and have not evolved significantly during the past three decades. To address the challenge of designing cellulose comminution processes for a reliable and predictable production of nanocelluloses, we engineered a study setup, referred to as Hyper Inertia Microfluidizer, to observe and quantify phenomena at high speeds and acceleration into microchannels, which is the underlying flow in homogenization. We study two different channel geometries, one with acceleration into a straight channel and one with acceleration into a 90 degrees bend, which resembles the commercial equipment for microfluidization. With the purpose of intensification of the nanocellulose production process, we focused on an efficient first pass fragmentation. Fibers are strained by the extensional flow upon acceleration into the microchannels, leading to buckling deformation and, at a higher velocity, fragmentation. The treatment induces sites of structural damage along and at the end of the fiber, which become a source for nanocellulose. Irrespectively on the treatment channel, these nanocelluloses are fibril-agglomerates, which are further reduced to smaller sizes. In a theoretical analysis, we identify fibril delamination as failure mode from bending by turbulent fluctuations in the flow as a comminution mechanism at the nanocellulose scale. Thus, we argue that intensification of the fibrillation can be achieved by an initial efficient fragmentation of the cellulose in smaller fragments, leading to a larger number of damaged sites for the nanocellulose production. Refinement of these nanocelluloses to fibrils is then achieved by an increase in critical bending events, i.e., decreasing the turbulent length scale and increasing the residence time of fibrils in the turbulent flow.

Fibre collection on screens is here used as a collective name for any preferred or non-preferred deposition of fibres, appearing in e.g. paper manufacturing, recycling of elongated particles or equipment clogging. The fibre collection on screens is typically discussed on basis of the application and may distinguish between fibre collection being a friend or foe of the process of interest. We report an extensive experimental investigation of fibre collection and provide a systematic discussion based on two parameters describing the screen geometry: the fibre length to the screen opening size (L-Fibre/D-Open), and fibre length to the distance between the openings (L-Fibre/S-Open). The first parameter, L-Fibre/D-Open, discriminates between the two fibre collection modes: (i) fibre retention (L-Fibre/D-Open, for example paper forming) and (ii) fibre stapling (small L-Fibre/D-Open, for example deposition on pins and edges). The second parameter L-Fibre/S-Open controls the fibre collection rate for both modes (with higher collection rates for higher values), but through different physics. In fibre retention, the successful collection is probabilistic and large for small screen openings and a fibre-orientation parallel to the screen. Since a decrease of L-Fibre/S-Open results in a smaller open area and hence to higher acceleration of the suspension upstream of the screen, the fibre orientation is skewed towards a screen normal orientation and fibres tend to pass through the holes. In the case of fibre stapling, the successful collection comes from an immobilization of the fibre when fibre-solid friction force exceeds the hydrodynamic drag force for fibres deposited close to the edge of the holes. For L-Fibre/S-Open above 1 a fibre bends over the solid hole spacing and is fixated on two support points. For L-Fibre/S(Open )below 1 restrains a successful fibre fixation on two support points why fibre collection on the solid screen is hindered and prevented. The impact of the approach velocity and fibre concentration on the fibre collection was tested and found to be negligible for fibre retention but impacting fibre stapling. This is in agreement with reports on equipment clogging in cellulose fibre processing. For fibre retention, the collection rate is high and close to total retention. However, the collection rate of stapling is much lower. At high velocities, we suggested that fibre bending can cause additional leakage through the screen.

Cellulose fibres are prone to flocculate and form aggregates that are deformable by the hydrodynamic stress. In this work I document for coiled pipe flow, known to have secondary motion in the pipe cross-sectional plane, an accumulation of fibre flocs and fibre aggregates at the outer bend. That is the segregation into a section in the pipe cross-section and hence presents a case of angular segregation. The segregation was studied for non-coherent crowded fibre flocs. For that, segregation benefited from fibre concentration and suffered from increased hydrodynamic stress expressed by increasing Reynolds number. Based on the observed segregation of fibres a flow splitter was designed that separated the flow at 1/3 of the tube diameter measured from the inner bend. The outer bend suspension length-weighted fibre length was found to increase. For the best case in this work, the difference between outer and inner bend relative to the feed fibre length was 22%. As for radial and axial segregation, which are known, also angular segregation is fibre-length sensitive. As such it can be exploited for length fractionation of networking and aggregating elongated particles, for example fibres. 

The motion of flocculated fibres in a streaming suspension is governed by the balance of the network strength and hydrodynamic forces. With increasing flow rate through a channel, (1) the network initially occupying all space, (2) is then compressed to the centre, and (3) ultimately dispersed. This classical view neglects fibres-fines: we find that the distribution of these small particles differs in streaming suspensions. While it is known that fibre-fines can escape the fibre network, we find that the distribution of fibre-fines is non-homogenous in the network during compression: fibre-fines can be caged and retarded in the streaming fibre network. Hence, the amount of fibre-fines is reduced outside of a fibre network and enriched at the network’s interface. Aiming on selectively removing fibre-fines from a streaming network by suction, we identify a reduction of the fines removal rate. That documents a hindered mobility of fibre-fines when moving through the network of fibres. Additionally, we found evidence, that the mobility of fibre-fines is dependent on the fibre-fines quality, and is higher for fibrillar fines. Consequently, we suggest that the quality of fibre-fines removed from the suspension can be controlled with the flow regime in the channel. Finally, we present a phenomenological model to compute the length dependent fibre distribution in an arbitary geometry. For a fibre suspension channel flow we are able to predict a length-dependent fibre segregation near the channel’s centre. The erosion of a plug of long fibres was however underestimated by our model. Interestingly, our model with parameters fitted to streaming fibre suspension qualitatively agreed with the motion of micro-fibrillated cellulose. This gives hope that devices for handling flocculated fibre suspensions can be designed in the future with greater confidence.