3D printing of a wood-based furniture element with liquid deposition modeling

Until now, mainly subtractive manufacturing processes have been used in wood processing. The material efficiency of these manufacturing technologies is naturally limited. Each year, the production of wooden furniture generates several hundred thousand tons of wood residues and chips. By using alternative technologies, the amount of wood residues could be reduced or used for the production of goods.

In additive manufacturing, material is deposited layer by layer based on a digital model and the process chain is therefore characterized by the direct production of components based on 3D CAD data. What all additive manufacturing processes have in common is that the material used remains completely in the finished workpiece or is available for the production of the next workpiece, which makes this process very material-efficient.

In the majority of cases, wood was used in additive manufacturing in combination with thermoplastics (Krapež Tomec and Kariž 2022). In the fused deposition modeling (FDM) process used for that purpose, the wood content is limited to a maximum of 40%. Liquid deposition modeling (LDM) represents an alternative 3D printing process. This technology can be used to process a wide variety of pasty materials: clay, loam, food, but also wood materials. LDM is much more energy efficient compared to other additive manufacturing processes (such as FDM, which requires energy to melt plastic) (Faludi et al. 2019). Materials are processed in the form of paste-like suspensions, with fillers (e.g., wood flour) dispersed as a solid phase in a liquid medium. Initial attempts to additively manufacture wood using LDM were described by Gardan and Roucoules (2014). Kariz et al. (2016) and Pitt et al. (2017) used urea-formaldehyde resins as binders and the overall wood content was less than 30%. Rosenthal et al. (2018) developed material mixtures for liquid deposition modeling with special focus on ecological aspects. Water-based binders from renewable resources were used (e.g., methyl cellulose, starch). The proportion of wood as a filler was increased up to 90%. The aim of an ongoing research project is to upscale the process. The present state of development will be briefly described in this paper.

First, a material mixture had to be developed with the aim of carrying out large-format tests. Carboxymethyl cellulose dissolved in water (sold under the trade name “Zell-Leim”, Baufan Bauchemie Leipzig GmbH, Germany) was used as binder and wood flour Lignobest^® C200 from Holzmühle Westerkamp GmbH (Visbek, Germany; particle size below 125 μm) formed the filler. The wood component is produced from softwood sawdust. As described in Rosenthal et al. (2018, 2023), the vertical material shrinkage during drying (n = 4), the flexural strength of the cured material (n = 5) and the density were determined.

A ScaraV4 3D printer from 3D Potter (Stuart, FL, USA) was used for the upscaling experiments (Fig. 1). The build volume has the geometry of a hollow cylinder. From a circular base area with a diameter of 1830 mm in the center a free space of 280 mm diameter (central axis of the 3D printer) is to be subtracted. The achievable height of the objects is 1320 mm. The objects were designed digitally with Fusion360 (Autodesk) and sliced with Simplify3d. A special vase mode was used as a printing strategy, which creates a continuous print path that forms a spiral in z-direction. The printing speed was 1800 mm/min, the layer height 5 mm.

The three mixture components binder (“Zell-Leim”), water and wood (Lignobest^® C200) were mixed in different proportions. The binder/water ratio of the tested mixtures was 1:10, 1:15 and 1:20, the binder/wood ratio 1:3 to 1:6. The wood content of the dried material varied accordingly, ranging from 75.0 to 85.7%.

The physical properties of the material are listed in Table 1. In vertical direction, the measurements showed a shrinkage of the test specimens of 13.3 to 19.5% due to drying. This aligns with results from other studies using entirely bio-based ingredients (Rosenthal et al. 2018; Sanandiya et al. 2018). Shrinkage decreased with increasing wood content for each of the three binder/water ratios. Rosenthal et al. (2023) showed that mineral additives can further reduce shrinkage significantly (but also increase the density, which is not desirable for many applications). High shrinkage values are problematic. They lead to deformation, cracking and delamination. They should not significantly exceed the value range of 15%.

The flexural strength ranged from 1.02 to 5.64 MPa. The density was between 0.36 and 0.48 g/cm³. A higher binder content and a lower wood content, respectively, resulted in improved mechanical properties and higher density values (R² > 0.92). In combination with certain binders, only low wood contents are possible; density and strength are accordingly significantly higher. Orji et al. (2022) developed mixtures with sodium silicate as binder. The wood content was 50 and 60%, and the density was 0.83 and 0.89 g/cm³, respectively. Flexural strengths of 25 MPa could be achieved. Pitt et al. (2017) reported flexural strength values of 57 MPa. However, they used urea resin as a binder, the wood content was only 13% and the cured material had a density of 1.22 g/cm³.

As a result of the dependence of both parameters on the binder or wood content, flexural strength and density also correlate very strongly (R² = 0.9). This does not only apply to the Carboxymethyl cellulose mixtures investigated. In other studies, this correlation was also found for methyl cellulose (Rosenthal et al. 2018, 2023) and chitosan (Sanandiya et al. 2018). By using elongated fibers, e.g., cellulose (Sanandiya et al. 2018) or thermomechanical pulp (Rosenthal et al. 2023), instead of shorter wood particles, it is possible within certain limits to enhance flexural strength without simultaneously increasing density values.

Three mixtures (1:10:3.5, 1:15:5, 1:20:6) showed noticeable cracking within the print paths. A reduced internal material cohesion due to a too high filler or too low binder content is assumed to be the cause. For further tests, the mixture 1:15:4.5 was selected. Among the mixtures without cracking, it was the one with a shrinkage below 16% and a flexural strength above 2 MPa. With a binder price of 4 €/kg and a wood price of 0.63 €/kg, the material cost was 1.24 €/kg.

For the production of small objects, piston extruders can generally be used (Kariz et al. 2016; Pitt et al. 2017; Rosenthal et al. 2018). The ScaraV4 was also originally equipped with a piston extruder (cartridge volume of about 4 L). The disadvantage of these discontinuous extruders is that they are not suitable for upscaling of liquid deposition modeling with wood-based materials, as they do not allow the production of larger objects. Changing cartridges resulted in visible misalignment and later delamination during the drying process at the layers, where the cartridges had to be changed. This problem could only be solved by a continuous printing process. The use of screw extruders is appropriate for this purpose (Dritsas et al. 2018; Orji et al. 2022). Therefore, a dual screw extruder was developed and experimentally manufactured from acrylonitrile butadiene styrene (ABS) using common FDM methods (Fig. 1): The horizontal, compressing screws have an always open feed zone for material intake between the screws and a closed compression section towards the extrusion nozzle. The material flow is redirected 90° in a flow-optimized manner to deposit the paste onto the layers through a 10 mm diameter nozzle. The screws are synchronously rotated by the printer’s stepper motor with a reducing gear. This enables both precise control of the extrusion and rapid blocking of the material outflow (retract) to prevent leakage during idle movement.

The pasty printing material was roughly pre-mixed by hand and filled into the vertical extruder chamber. Homogenization of the wood paste and evacuation of the air took place in the extruder. A 10 mm thick chipboard sheet covered with a stretch foil (packaging material) served as a print bed. After the printing process, the foil was cut off around the printed object. This allowed uniform shrinkage of the drying objects. The drying process was enhanced by three cold (room temperature) air blowers (hot air leads to the formation of cracks). To prevent mold growth, air movement over the entire object surface had to be ensured. After about 5 days of drying, the objects were ready for transport. It takes a total of 7 to 14 days to reach the final strength. The material mixture of carboxymethyl cellulose and wood flour used is not resistant to moisture. By adding water, it can be converted back into a paste-like substance and used for the next print (this allows recycling). If moisture resistance is desired for a specific application, additives have to be added to the mixture or a surface treatment has to be carried out subsequently.

As a demonstrator of the object geometries achievable so far, the frame of a small side table (base 526 mm x 526 mm, height 180 mm, dry mass 640 g) was manufactured (Fig. 1). The design concept was inspired by Islamic ornamentation. Curvatures improved wall stability during the printing process. Cylindrical structures at the frame corners subsequently served to hold the table legs.

Liquid deposition modeling with wood-based materials is a suitable technology for fostering a digital and at the same time sustainable transformation of furniture production. As part of ongoing development activities, the technology readiness level has been increased from 4 (lab-scale) to 5 (pilot-scale). Future research efforts should be primarily oriented towards the following goals: 1. Drying-related shrinkage in the double-digit percentage range represents a challenge for object design and the drying process. New material mixtures should have lower shrinkage if possible. 2. The material strength of the mixture used is sufficient for the production of furniture elements that are combined with other materials (e.g. freely formed edge elements of upholstered furniture frames). Improved materials need to be developed for more mechanically stressed applications. 3. Further research is also needed with respect to object dimensions. Currently, objects with a height of up to 20 cm can be produced. Higher objects collapse during the printing process or shortly thereafter. The instability of the wet mass also limits design freedom in the vertical direction. Geometries with overhang are very challenging. These problems can only be solved by highly accelerated curing times of the material. 4. For production on an industrial scale, automation of all process steps is required, especially mixing and conveying the material to the extruder. Solving the aforementioned challenges means achieving important milestones on the way to practical application.