Three‐dimensional forming of plastic‐coated fibre‐based materials using a thermoforming process

Three‐dimensional (3D) forming of fibre‐based materials has been a topic of growing interest over recent years and 3D forming processes using hydroforming, press forming and deep drawing processes have been widely explored. Thermoforming as a potential alternative method for forming these materials remains, however, relatively understudied. This research attempts to provide a fundamental understanding of the thermoforming limitations of plastic‐coated paperboards. In the work, a variety of commercial paperboards are subjected to experimental tests with different forming parameters and moulding depths. Shape accuracy, maximum acquired depth, thickness distribution behaviour and damage mechanisms are used to evaluate thermoformability, and the results linked to the material properties and forming conditions. The research findings indicate that the plastic‐coated paperboards studied are thermoformable but only in simple geometric shapes and with low mould depths. Unlike plastic, thermoforming can result in thickness increase in plastic‐coated paperboards, which is thought to be a result of out‐of‐plane auxetic behaviour of paperboards. Paperboard thermoforming was also found to be hindered by rupture, blistering and curling defects. Tensile strain at break is the key factor determining thermoformability. Additionally, the density of the paperboard can impact the heating step and the rate at which the moisture content of the material changes during the forming process. Furthermore, it was observed that changes in process parameters affected materials differently, with the direction and rate of change differing based on the material being used.


| INTRODUCTION
Increasing awareness of the importance of environmental sustainability has led to widespread recognition of the need to reduce global use of plastic, which is a non-biodegradable, oil-based product, and invest in sustainable alternatives. Fibre-based composites are known as a possible alternative to plastic; while retaining the beneficial properties of plastic, they contain a high proportion of biodegradable fibrous material, resulting in a lower climate impact. As fibre-based products begin to replace plastic ones, improvements are needed in the area of conversion of these materials into versatile shapes. Therefore, 3D forming of fibre-based materials has been a topic of increasing research extensively used in the plastics industry, where its high production speed, low pressure and temperature requirements for forming customized products make it a cost-effective and popular manufacturing method. 17 In the area of the production of fibre-based products, however, thermoforming has received little research attention. The lack of work can partly be attributed to the nature of the process, which largely matches the properties of plastic materials, and as noted by Östlund, 18 the challenges posed by the porous structure of fibrebased materials. Thermoforming of fibre-based materials has thus far been researched for relatively flat moulded part products. 19 Some work, also, has been done in the area of thermoforming of fibre-based composites in mixed structures, such as wood polymer composites (WPC), where the focus has been on thermoformability investigations using finite element methods. 20,21 To date, thermoforming of plastic-coated fibre-based materials has been relatively understudied. Although the plastic coating allows these materials to be thermoformed, thermoforming of plastic-coated fibre-based materials faces certain limitations. Therefore, the primary objective of this research work is to gain a deeper insight into the thermoforming limitations of plastic-coated fibre-based materials by comparing the results for such materials with those for full plastic materials. Additionally, the effects of material properties and three key process parameters, that is, forming temperature, forming time and forming pressure, on forming results are explored. The knowledge gained can provide a foundation for future research investigating process and product development.

| Materials
Two types of materials were used in this research: multilayer plastic and plastic-coated packaging paperboard. APET/EVOH/PE multilayer plastic was selected as a reference material due to its excellent thermoformability, which allows the authors to fully demonstrate the thermoformability limitations of coated fibre-based materials. The paperboards selected for study were commercial polyethylene (PE)and polyethylene terephthalate (PET)-coated boards from various suppliers to enable analysis of their thermoformability differences. Plastic material, as named Ref in Table 1 The moisture content of the paperboard materials was measured using the moisture analyser (Adams Equipment PMB 53, New York, USA). Upon putting 3 g of the material into the moisture analyser, the material is gradually heated until the moisture is completely evaporated. The moisture content in a material is determined by the difference between its initial and final weight.
Paperboard 1 consists of one layer of paperboard sandwiched between two plastic layers. In order to ensure that moisture content on that layer is equally accounted for in measurements, the material is delaminated from the sandwiched paperboard layer (Figure 2), allowing moisture to fully evaporate during measurements from both paperboard layers. The thermoformability of the materials was examined using variation of forming parameters and two thermoforming processes: vacuum thermoforming and a combination of vacuum and air-pressure thermoforming ( Figure 5). During forming, the pressure was controlled directly using an air-pressure valve. A pressure sensor was used to indicate the actual value of the forming pressure in the forming chamber. Various different mould depths were used to assess each material's capabilities in terms of achieving the shape depth. The range of process parameters investigated in this study is presented in Table 2.
The forming parameters were chosen on the basis of the results of preliminary experiments and the limitations of the machine, and the materials being formed. It should be noted that since studying the effects of four parameters at three levels by examining the interaction between them is time-consuming and costly when five different materials are used, this study concentrates on testing the effects of each parameter individually.

| Thermoformability analysis
Several factors must be taken into account when evaluating the formability of different materials, and no consensus exists regarding the right criteria for assessing formability. In this research, shape accuracy, maximum depth acquired, wall thickness distribution and damage mechanisms are considered indicators of material formability. Prior to the measurements, all the samples were conditioned at 23 C ± 1 C and constant humidity of 50% ± 2% RH. To ensure the inclusion of postdeformations and springbacks into the analysis, measurements were performed a few days after the experiments. For each set of parameters, measurements were repeated for six identical samples and average values were used to derive the results.
The shape accuracy was first assessed by visual inspection, and the samples were then analysed with the 3D measurement system (Keyence VR-3200, Osaka, Japan). This analysis was done to gain a better understanding of shape accuracy differences between the various materials and to provide quantitative values for this quality factor.
In the evaluation, the defined area shown in Figure 6 was chosen and a 3D model generated. The CD profile was used for further measurement. Next, the profile depth and the corner angle derived from the side lines of the profile were measured; these amounts were, respectively, 12 mm and 60 in the mould design. In order to perform this 3D measurement analysis, the formed plastic samples were painted with a zinc colour to make them clearly visible to the measurement system and to overcome issues caused by reflections. Moreover, due to limitations of the system in the depth of analysis, the plastic sample was analysed in two steps, and the results were merged together.
The forming depth of all the samples was measured using a hand- Furthermore, damage mechanisms that can further hinder the forming of materials or restricts their use were identified through visual inspection. In this context, the plastic coating layer of the material was also examined. Addition of plastic coatings to paperboard has the primary function of enhancing the barrier properties of the material.
Hence, this study also examines whether thermoforming has an adverse effect on the coating layer of the studied paperboards and its barrier performance. Using the dye penetration test, it was determined if thermoforming causes the formation of pinholes in coating layer and deteriorates its barrier properties. The experiment was conducted in which a colouring solution was applied to the coating side of the thermoformed tray, and the other side of the tray was inspected for penetration. The tests were conducted according to EN standard 13676 (2001) using a colouring solution consists of 0.5 gram of dyestuff E131 blue dissolved in 100 ml of ethanol (96%). 24 Moreover, the SEM imaging was also used to investigate the changes in the coating layer of the materials after thermoforming. Prior to forming, materials were imaged from their coating side; later, thermoforming trays were also imaged from their coating side.

| Optical analysis
The optical analysis carried out throughout this study was performed using the SEM Hitachi SU3500 (Tokyo, Japan), equipped with a tungsten filament. The plastic coating side of the materials was imaged  for the range of process parameters investigated is given in Figure 10.
It should be noted that the geometry of the mould used can have a significant effect on the maximum depth that can be achieved by the material; the maximum depth may thus vary when using another geometry. To find the extent at which each of the forming parameters affected the attainable depth for each material, statistical analysis with Minitab software was carried out ( Figure 11). Based on multiple regression coefficient analysis, it can be concluded that increasing forming pressure is more likely to enable Paperboards 1, 3 and 4 to achieve greater depth. For Paperboard 2, however, the forming temperature affects the material more significantly. The higher effect of forming temperature on Paperboard 2 is further explained in Section 3.2, which discusses density effect on thermoformability of paperboard materials.

| Thickness distribution
It is important to investigate thickness distribution in thermoformed products. Generally, the stretching in the process reduces the material's thickness compared to its raw state, and if the process is not controlled properly, the material can become thin, which can significantly degrade the quality and strength of the formed part. Similarly, uniform distribution of thickness is essential for good quality products. Accordingly, in this study, thickness was measured over the cross-sectional midplane profile of the selected samples indicated in The most likely explanation for the increase in the thickness for the Paperboards 2 to 4 can be attributed to a phenomenon known as out-of-plane auxetic behaviour of paper. When paper is stretched under tensile deformation (e.g., during thermoforming), thickening can occur with the stretching of the fibres, and this phenomenon has been termed auxetic behaviour. 27,28 Fibre networks are characterized generally by curled fibres, while some fibres are found on top of each other and some below. Stretching a material causes the curled fibres to straighten, and with this change, the fibres on top of them are pushed up, resulting in the thickness increasing. 27 In the example of this study, thickening occurred in Points 2 to 6; the reason is because the material starts to be stretched where the depth is changed as it is the case from Points 2 to 6. In addition, increasing in the thickness was greater on side walls compared to the bottom. This further suggest that rate of stretching is greater on side walls compared to the bottom. This could be attributed to the curved shape of side walls compared to the flat bottom found on paperboard samples ( Figure 12). Other studies have also found an increase in the thickness of paper-type materials under tensile deformation. 29, 30 Verma et al. 27 further demonstrated that the rate of this auxetic behaviour can vary depending on the structure of fibre networks and the method of preparing paper materials. Similarly, the rate of thickness increase between Paperboards 2 and 4 is different in the current study, as shown by Figure 12. The differing performance of Paperboard 1 could not be explained clearly, but its layer structure may partially account for its behaviour. In addition to the coating layer, Paperboard 1 has another plastic layer incorporated between layers of paperboards, which may cause the material to behave more like plastic. However, a further investigation regarding the mechanism behind this difference remains to be conducted.
Regarding the effect of the process parameters on the thickness distribution, the experiments indicated that in contrast to the forming depth, changing the forming parameters within the specified range had little effect on the thickness distribution. Statistical analysis of the results also showed that changes in the forming parameters had no statistically significant influence on the thickness. Hence, at the studied range of parameters, the elongation properties of the materials tend to dominate in thickness distribution of the coated paperboards and optimizing the process parameters has little influence.

| Tensile properties
Tensile properties can be considered one of the most important properties of materials for thermoforming. The formability of fibre-based materials is dependent on various mechanical properties, for example, elongation, compressive strain, compressive strength and substrateto-metal friction, and the degree to which these properties cause changes in the formability varies widely depending on the forming equipment and forming parameters. 1 Unlike press forming and deep drawing, thermoforming displays tensile deformation rather than compressive deformation; and therefore, elongation is the most critical of the above-mentioned properties. Basically, elongation is a measurement of extensibility, which is the degree to which the linear length of a material can be extended by elastic, viscoelastic or plastic deformation when external forces are present, and extensibility in tensile deformation is related to strain at break tensile properties. 31 Tensile tests indicated that Paperboard 1 has a sixfold higher strain at break value in the MD and a two-fold higher value in the CD than the other paperboards, which may partly explain its better shape conformance. Considering the forming angle in the deep grooves and the maximum achieved depth of the trays, correlation analysis indicates a strong relation between the CD strain at break properties of F I G U R E 1 2 Mapping of thickness points on a sample of Paperboard 1 F I G U R E 1 3 Thickness distribution considering the thickness of raw material as zero the paperboards and their ability to form the mould shapes ( Figure 14). According to the study by Östlund et al., 32 the strain at break at CD is a dominant factor in the 3D formability of paper materials. The same trend is seen in the MD direction with correlation values of À0.86 and 0.80 for the forming angle and forming depth, respectively. Moreover, looking at the thickness distribution presented in Figure 13, and excluding the edges where no stretching occurred, it can be seen that the stretching of the materials is correlated with their strain at break properties. Based on the change in thickness, greater strain at break results in greater stretching ability. In Figure 14, this correlation can be seen in the CD direction, and correspondingly, the correlation value is 0.90 in the MD direction. It is nevertheless important to note, as demonstrated by the R 2 values, that the strain at break properties and formability factors do not follow an exact linear relationship because of the difference between uniaxial and 3D strain behaviour. However, these results suggest the higher strain at break properties are beneficial to improve thermoforming performance while the rate of change in thermoforming performance cannot be completely predicted using the uniaxial strain at break properties of the materials.

| Density
Thermoforming using roll-fed machines involves heating the material prior to forming. Thermoforming is generally used to form plastic materials, and this preheating step has the benefit of drying out moisture that has been absorbed by the plastic and is detrimental to its forming process. In fibre-based materials, however, the existing moisture has a crucial impact on the forming capability. The presence of moisture in paper material can make it softer and improve its formability as the greater its moisture content, the weaker the fibre bonds, and so the paper has a lower elastic modulus and tensile strength, making it easier to stretch and form. 2,33 Consequently, to better understand the performance of paperboard materials in thermoforming, the changes in moisture content during the forming were investigated for the specified forming conditions (110 C forming/heating temperature, 3 s forming/heating time and vacuum plus 1 bar forming pressure). The moisture content of the material was measured immediately before forming from the material roll and directly after forming from the formed sample. The rate of change in moisture content of the materials was calculated accordingly and presented in Table 3.
Using the Fourier number, which is the ratio of heat conducted through a material to heat stored during the heating process, it can be possible to explain the rate of changes in the moisture content of the materials: Fourier number ¼ Thermal diffusivity of the material Â Heating time In the experiments in this study, the heating time was constant for all the materials, and the material thicknesses are fairly close; thus, the thermal diffusivity of the materials can be considered the determining factor. The thermal diffusivity of a material is an indicator of how rapidly its temperature changes when heated, and according to heat transfer laws, it is inversely related to the density of the material.
Nevertheless, there has been disagreement regarding the dependency of thermal diffusivity on the density of paper materials. In contrast to the research conducted by Niskanen and Simula, 34 which showed that the thermal diffusivity of paper material is inherently independent of its density, Morikawa and Hashimoto 35 found an inverse relationship between these two variables. The reason for this is explained by the structure of paper materials, which are made up of cellulose and air; the higher density corresponds with fewer pore spaces and lower air content in the material. Air has a much higher thermal diffusivity than fibres, so the higher the density, the less air is present in the material, which translates to a lower thermal diffusivity. 35 Accordingly, in this study as well, the density of paperboard layers is calculated to determine whether the discussed inverse relationship is valid when considering the changes in moisture content in materials.
F I G U R E 1 4 Correlation between the strain at break properties of the paperboards and their formability factors (forming angle, forming depth and maximum thickness changes) To calculate the density of paperboards in the studied materials, it is necessary to know the thickness of the paperboard layer. As such, the cross-sectional SEM images of materials are used to determine the thickness of paperboard layers. The ratio between the thickness in the paperboard layer and the total thickness is determined by using the available length measurement tools for images (in this study with the help of Microsoft Visio). Accordingly, the paperboard layer's thickness is calculated based on the measured total thickness of material, and its density is calculated by dividing the grammage of paperboard (shown in Table 1) by its thickness. As an example, Figure 15 demonstrates how the thickness of the paperboard layer of Paperboard 2 was determined. This method may not be able to provide an exact thickness of the layer due to variations in thickness, but it can provide a fair indicative value.
To assess the effect of density on heating of the material and the potential relationship with the change in moisture content, Paperboard 1 has been excluded from the study. This is due to the differing layer structure of this material in comparison to the three other paperboards as seen in Figure 1. Unlike the other three paperboards, this material has an internal plastic layer between two paperboard layers, resulting in a different mechanism of distributing heat. Moreover, to compare this material with other materials, density of each paperboard layer is needed for the calculation, which requires information regarding the grammage of each layer and that information is not available in this study. Thus, to provide a valid comparison, the relationship between density and moisture content change is examined between the three other coated paperboards.
The results of this study show that there is a strong correlation between the moisture content changes in material during thermoforming and their density ( Figure 16). Therefore, when comparing all the materials for the same heating time, Paperboard 2 with the lowest density had the highest thermal diffusivity and rate of heat conduction, as well as the highest temperature increase, leading to a higher change in its moisture content. These results suggest that using denser paperboard materials can aid in moisture retention during the thermoforming process.
Conversely, however, when it is claimed that increasing density leads to better moisture retention in the material due to lower temperature rise, it might be argued that at the same time higher temperatures are needed to enable the material to be formed better. In this study, there is no information pertaining to the difference between the set heating temperature and the actual material temperature.
Nevertheless, it can be assumed that with same heating time, the denser paperboard had lower actual temperature or that there was an uneven distribution of temperature in the material, which would account for the surfaces of the material reaching the required temperature, while the cores of the material did not reach the same temperature. Interestingly, this assumption is in accordance with the effect of forming temperature on the depth of trays. From Figure 11, it is apparent that the less dense material is more affected by forming

| Damage mechanisms
The most common defect observed in the thermoforming of the paperboard materials was rupture. The mode of rupture in the thermoforming experiment is shown in Figure 17, where the fracture occurs along the longitudinal MD. Defects such as rupture are caused by stress levels exceeding the strength of the material. The bonds within the fibres fail first, followed by the fibres themselves failing. 36 In the thermoforming process, tensile deformation can also result in tensile failure and fracture formation. Rupture occurrence depends on both the material properties and process parameters. In this present study, rupture was most frequently caused by an increase in the mould depth. The moisture content of the material was also observed to be a very significant factor as low moisture content resulted in earlier rupture as the depth increased. Yet, in some cases, a drastic increase in forming temperatures could also cause rupture.
Blistering was another observed defect. Blistering can be caused by excessive heating of the material, as well as excessive moisture content where the water vapour cannot leave the uncoated side of the paperboard, which results in internal delamination and blistering.
Blistering can be prevented by selecting appropriate processing parameters and applying a barrier coating with high adhesion to the paper. 2 All studied paperboards showed blistering when formed at high temperatures, although these temperatures ranged according to the material. An example of this defect can be seen in Figure 17, which depicts blistering on Paperboard 1 after the forming temperature was increased to 140 C.
Curling was a defect that occurred during storage. Curl, as defined by Salmen, 37 is a specific type of dimensional instability which occurs when the paper material is deformed over its elastic limit.
According to Niini et al., 38 this phenomenon has been also seen in press-formed products, and it is primarily due to hygroexpansion, stress relaxation and creep of materials while absorbing or desorbing moisture in different humidity conditions. In this study, in spite of having stored all the samples in the same condition, samples made of Paperboard 1 exhibited curling during storage. The higher plastic content of this material and its considerably better elongation properties make the material less stiff and more flexible, which is very useful in terms of forming but at the same time can cause curling and instability of the formed product during storage. Further investigation is F I G U R E 1 7 Rupture, blistering and curling in the thermoformed plasticcoated paperboard samples F I G U R E 1 8 SEM images of coating layer for Paperboard 3 before and after the thermoforming process required to ascertain the ideal storing conditions to avoid curling in such fibre-plastic materials.
The dye penetration test was conducted on samples obtained under different forming conditions to analyse the potential damage to the plastic coating layer of the studied paperboards. The results showed that the thermoformed trays studied did not leak regardless of their type of coating (PE or PET) and the conditions under which they were formed. This can generally verify that the thermoforming process does not cause pinholes to appear in the coating layer of the materials.
In addition, SEM imaging of thermoformed trays showed no surface defects after forming. As an example of this, Figure 18 illustrates SEM images of the coating side of Paperboard 3 before and after forming.
Images also depict a smoother coating layer after the forming, which itself could be beneficial in possible subsequent sealing processes of thermoformed trays. Overall, this study found no evidence of pinholes or surface defects in the plastic coating of the thermoformed trays studied. Nevertheless, a more detailed investigation of the barrier properties of thermoformed trays remains to be conducted.

| CONCLUSIONS
This study evaluated the thermoformability of plastic-coated paperboards based on their ability to replicate mould shapes, the maximum achievable depth, thickness distribution behaviour, and damage mechanism. Further, the effect of material properties and process parameters on formability was discussed. A plastic material with excellent thermoformability was used as a reference to better visualize the limitations of the coated paperboards.
The thermoforming of paperboard is greatly restricted by the nature of the material, which contributes to its low stretchability. Plastic coating provides a certain thermoforming window to produce 3D products from these materials without causing structural damage.
However, in comparison to fully plastic material, the thermoformability of plastic-coated paperboards is limited to simple geometric forms and shallow depths. Indeed, thermoforming of paperboard yields low-depth balloon-shaped products rather than the more complex geometrical shapes such as patterned-shapes and deep grooves that are possible when using plastic. In addition, although thermoforming tackles the issue of possible thickness reduction and thinning that is found in plastic materials, coated paperboards can undergo a reverse trend where there is an increase in thickness.
Thickening is attributed to the out-of-plane auxetic behaviour of paperboard materials. Furthermore, damage mechanisms such as rupture and blistering pose further limitations to the thermoforming of paperboards, while curling disrupts post-usage of products. On the other hand, the plastic coating of the thermoformed trays tested showed no signs of pinholes or surface imperfections.
Empirical study and subsequent statistical analysis revealed that careful selection of the plastic-coated paperboard material is required to achieve acceptable thermoforming performance, with attention to be given to the tensile elongation and density properties. The higher the strain at break of the material, the deeper its attainable forming depth, and the better its stretch and shape conformity. Additionally, the higher density of paperboard results in less thermal diffusivity and a lower rate of heat conduction. Consequently, paperboard of higher density can experience lower moisture content changes for the same heating time than less dense paperboard. Correspondingly, however, it requires longer heating time to reach the required forming temperature.
The results of this study show that no simple change, increase or decrease, in the forming parameters has the same effect on all materials, and the degree to which parameter changes affect the materials varies. Therefore, the process conditions and tooling require specific adjustment to the material used to provide good thermoforming performance and yield defect-free products of high quality.