**Formaldehyde Emission in Micron-Sized Wollastonite-Treated Plywood Bonded with Soy Flour and Urea-Formaldehyde Resin**

**Hamid R. Taghiyari 1,\* , Seyed Behzad Hosseini <sup>2</sup> , Saman Ghahri <sup>2</sup> , Mohammad Ghofrani <sup>1</sup> and Antonios N. Papadopoulos 3,\***


Received: 7 September 2020; Accepted: 23 September 2020; Published: 25 September 2020 -

**Abstract:** Soy flour was partly substituted for urea-formaldehyde (UF) resin with different content to investigate its effect on formaldehyde emission in three-layer plywood panels. In each square meter of panels, 300 g of resin was used (wet weight basis of resin). Micron-sized wollastonite was added to the resin mixture at 5% and 10% consumption levels (wet weight basis of resin) to determine its potential effects as a reinforcing filler to mitigate the negative effects of addition of soy flour. Results showed a decreasing trend in formaldehyde emission as soy flour content increased to 20%. The highest shear-strength values were observed in panels with 10% and 15% soy flour content. The addition of wollastonite did not have a significant effect on formaldehyde emission, but it decreased the shear strength in soy-treated panels, although the values were still higher than those of control panels. Wollastonite significantly mitigated the negative effects of soy flour on the water absorption and thickness swelling of panels. It was concluded that 10% of soy flour and 5% of wollastonite provided the lowest formaldehyde emission and the most optimum physical and mechanical properties.

**Keywords:** biobased resins; formaldehyde emission; minerals; wollastonite; wood composite panels

#### **1. Introduction**

Adhesives play an important role in the efficient utilization of wood resources and in the development and growth of the forest product industry. Adhesive bonding of solid wood and wood particles of various sizes is a key factor for the production of modern, functional wood products, used in a variety of applications. For centuries, wood was bonded using biobased adhesives until synthetic adhesives, mainly thermosetting ones, gradually took over in the 20th century, as they were typically regarded as more effective, cost-efficient [1], and stable for use in humid conditions. Today, the main classes of thermosetting adhesives are amino-based, phenolic, and isocyanate resins. The utilization of these thermosetting adhesives is considered more economical, and reactive adhesives with quick curing behavior are versatile in a range of properties in the cured state. These adhesives have dominated the wood composite industry for many decades. Within this group, urea-formaldehyde (UF) resins are the most important adhesives in terms of quantity. Due to their low-cost raw materials, their rapid curing, their high dry-bond strength, and a colorless glue line, UF-based adhesives are almost exclusively used for producing wood-based materials, such as particleboard or medium-density fiberboards

for interior applications [2]. When products are utilized in conditions exhibiting higher humidity, UF resins are usually modified with significantly more expensive compounds such as melamine, phenol, or resorcinol [2]. It has to be mentioned that the final adhesive composition in use depends on the requirements of the wood-based material such as the required strength properties, the expected moisture resistance, the production cost of the finished product, and the desired formaldehyde emission class.

When it comes to formaldehyde emission from wood-based composites, it has to be mentioned that emission can originate from (i) synthetic-free formaldehyde that is not polymerized into the network, which emits during or quickly after panel production, (ii) formaldehyde released due to adhesive hydrolysis, which emits over the lifetime of the panel depending on moisture and temperature, and (iii) biogenic sources [2]. The discussion surrounding formaldehyde started as early as the mid-1960s, as reviewed by Roffael et al. [3], and various stages of reduced emissions were achieved in the 1970s and intensified in the 1980s when the carcinogenicity of formaldehyde in rats and mice after long-term inhalation exposure was reported. The topic of formaldehyde in indoor environments was intensively and controversially discussed by various authors, as reviewed by Salthammer et al. [4]. Driven by the standard requirements specified by the local authorities in, e.g., Europe, Japan, and the United States of America (USA), formaldehyde content and emissions from wood-based composites have been continuously reduced over the last few decades.

During the last few decades, the wood industry made great effort with many innovations in amino resin technology to gradually reduce formaldehyde emissions in order to fulfill the individual product and standard requirements for each type of composite. These can be summarized as follows: (i) low-formaldehyde-content resins, (ii) formaldehyde scavenger additives, (iii) post treatments, and (iv) alternative adhesives [1].

For the first type of approach, the first tendency has been to prepare engineered UF resins of progressively lower molar ratio, at levels much lower than 1:1 [5], which has become rather common in industry today. This is due to the attempt to minimize formaldehyde emissions from wood panels bonded with UF resins. One of the drawbacks of the much lower than 1:1 molar ratio has been identified in an increase in the tendency of the UF resin to form increasingly present crystalline domains upon hardening as a result of hydrogen between linear molecules [6–8]. At higher molar ratios, the hardened resin is amorphous, affording better adhesion and better bonding performance. The overly high crystallinity drawback was very recently solved [9] by blocking the formation of hydrogen bonds using transition metal ion–bentonite nanoclay through in situ intercalation and, thus, converting the crystalline domains of the UF resins to amorphous polymers. Addition of 5% nanoclay to the UF yielded in excess of 50% better adhesion and almost 50% lower formaldehyde emission, thus resulting in a marked improvement in performance with a low level of crystallinity. In the same trend, the potential introduction of a very acidic pH condensation step in the preparation of UF resins, inducing the formation of occasionally considerable amounts of uron (a cyclic intramolecular urea methylene ether) in the UF resins with lower formaldehyde emission, has attracted some research interest [10,11]. This initial work indicated that introduction of such an acid step can lead to UF resins with improved bonding strength, but also higher post-cure formaldehyde emission.

With regard to the second approach, formaldehyde scavengers, capable of capturing formaldehyde either physically or chemically and forming stable products, are added to UF resins or to wood particles before pressing [12,13]. These additives should provide long-term formaldehyde emission reduction, in principle, along the panel's life. Examples used in industry include the addition of urea in an aqueous solution or powder form, organic amines, scavenger resins, sulfites, functionalized paraffin waxes, and porous absorbers such as pozzolan and charcoal [14]. Very good formaldehyde emission reduction is obtainable by adding sodium metabisulfite to the resin, by adding tannin solution to urea-formaldehyde resin, or by using different starch derivatives [15,16].

The third approach involves treatments that are applied after pressing. Currently used methods include panel impregnation with formaldehyde-scavenging species, such as aqueous solutions of ammonia, ammonium salts, or urea. Another option is the creation of diffusional barriers in the panel surfaces that keep formaldehyde confined, by using paints, varnishes, veneers, laminates, or resin-impregnated papers [17,18].

Alternative adhesives involve isocyanate-based adhesives and biobased wood adhesives. The concern about formaldehyde emission vapor levels from UF adhesives has brought isocyanate adhesives to the fore, where formaldehyde emission does not occur as no formaldehyde is added. pMDI (polymeric methylenediphenyl isocyanate) is an excellent adhesive and can be used in markedly smaller proportions than formaldehyde-based adhesives to bind wood composites. Another attractive option is biobased adhesives. The term biobased adhesive has come to be used in a very well specified and narrow sense to only include those materials of natural, nonmineral origin which can be used as such or after small modifications to reproduce the behavior and performance of synthetic resins. Thus, only a limited number of materials can be currently included, at a stretch, in the narrowest sense of this definition. These are tannins, lignin, carbohydrates, unsaturated oils, proteins and protein hydrolysates, dissolved wood, and wood welding by self-adhesion. An excellent review on this topic was recently published by Pizzi et al. [1]

The most common biobased adhesives are protein-based sourced from animal bones and hides, milk (casein), blood, fish skins, and soybeans. Soy protein is obtained from soybean and has been used for centuries as a wood adhesive. In the context of wood composite production, soy protein was added to phenol formaldehyde resin to lower formaldehyde emission, but lower water resistance is an important limitation [19]. Formaldehyde-free wood composites were obtained using an adhesive based on soy flour and glyoxal, a nontoxic, but less reactive, aldehyde [20]. It was reported that the use of soy protein combined with polyamidoamine-epichlorohydrin (PAE) resins yields a strong and water-resistant product that is commercially available for wood composites [21]. Another interesting formaldehyde-free adhesive system, successfully tested in the production of plywood and OSB (Oriented Strand Board) panels, is based on a combination of soy flour, polyethylenimine, maleic anhydride, and sodium hydroxide [19]. Hosseini et al. [22] reported that the partial replacement of urea-formaldehyde adhesive with soy flour, particularly with a substitution rate of 15%, significantly reduced the formaldehyde emission, while it did not significantly influence the shear strength, under both dry and wet conditions. Kawalerczyk et al. [23] applied five types of flours (rye, hemp, coconut, rice, and pumpkin) as fillers with urea-formaldehyde resin in plywood manufacture. It was reported that the type of flour had a major influence on the properties of resin mixture such as gel time, solid content, and viscosity. The use of hemp flour as a filler led to a substantial decrease in free formaldehyde content.

Under this context, the present study was carried out to primarily investigate the effects of partial substitution of soy flour for UF resin and the consequent effects on formaldehyde emission and on some key physical and mechanical properties. In order to minimize possible negative effects on properties due to the use of soy flour, an innovative approach was made in an extra set of boards, to add micron-sized wollastonite. Micron-sized wollastonite is a mineral that was successful in improving the physical and mechanical properties in medium-density fiberboards and particleboards [24–28]. It significantly improved the shear strength of polyvinylacetate resin [29]. It was even effective in improving the fire properties in wood-based composite panels and solid wood [29–35]. Therefore, this study continued with the addition of micron-sized wollastonite to the resin mixture to determine its potential effect on the properties of plywood panels produced with a mixture of UF resin and soy flour.

#### **2. Materials and Methods**

#### *2.1. Panel Production*

Three-layer plywood panels were produced, using poplar veneer (*Populus deltoides*) with 2.1 mm of thickness. The dimensions of the produced panels were 350 × 350 mm. The target thickness of the panels was 6 mm. For each square meter of panels, 300 g of urea-formaldehyde resin was used (wet weight basis of resin). Glued veneers were hot-pressed for 5 min at 130 ◦C. The specific pressure of plates was 1.5 MPa, and the total nominal pressure of the plates was 20 MPa. Once produced, 25 mm of each side was trimmed to avoid inconsistent edges (Figure 1). Except for the formaldehyde emission tests, all produced panels were kept in a conditioning chamber (25 ± 3 ◦C, relative humidity 60–65%) for a week before test specimens were cut to size. The specimens were kept under the same conditions for two more weeks before tests were carried out. The target density of the panels was 0.55 g·m−<sup>3</sup> . <sup>−</sup>

**Figure 1.** Panels produced and trimmed, ready to be cut to size for each test according to the relevant standard (**A**); front surface of a trimmed plywood panel (**B**).

#### *2.2. Resin Application*

UF resin was purchased from Amol Resin Company (Amol, Iran). The viscosity of the resin was 200–400 cP, with 47 s of gel time, and a density of 1.277 g/cm<sup>3</sup> . Defatted soy flour (SF) was purchased from Behpak Company (Behshahr, Iran). SF contained 47% (*w*/*w*) protein (Ghahri et al. 2016). SF was substituted for UF resin at 5%, 10%, 15%, and 20% (*w*/*w* dry weight). SF and UF were mixed for 10 min using a magnetic stirrer. Micron-sized wollastonite was prepared by Mehrabadi Machinery Mfg Co. (Tehran, Iran) (Table 1). Wollastonite was mixed with the resin for 10 min at 5% and 10% (*w*/*w* dry weight basis of resin). Once the resin was prepared, it was applied onto the veneers in less than 1 min. Just before applying the prepared resin on veneers, 1% ammonium chloride was mixed as a hardener (dry weight basis of UF resin). Within 4 min of the application of resin on veneers, the layers were arranged and then hot-pressed.



#### *2.3. Measurement of Formaldehyde Emission*

Formaldehyde emission (FE) in the present study was measured on the basis of European standard specifications (EN 717-3/Part 3) [36]. From each of the five replicate panels, three specimens were cut for the formaldehyde emission test. Necessary coordination was made so that specimens were tested immediately to minimize sources of error. The dimensions of the FE test specimens were 25 × 25 × 6 mm. Flask type 2 (with a volume of 500 mL) was employed. FE tests were carried out at the temperature of 40 ± 2 ◦C, for a duration of 180 min. In this test method, FE specimens were hung in vertical position at 40 mm above the distilled water (50 mL) at the bottom of the flask (Figure 2). The flask was then cooled (using a mixture of ice water) for 30 min to ensure absorption of the emitted formaldehyde in the distilled water inside the flask. Acetylacetone spectrophotometric analysis was used to determine the amount of formaldehyde in each flask. The determination of formaldehyde emission was based on the Hantzsch reaction. In this method, aqueous formaldehyde reacted with ammonium ions and acetylacetone to yield diacetyldihydrolutidine (DDL). DDL has a maximum absorption capacity at wavelength of 412 nm. A T60 visible range spectrophotometer was used in the present study, with a fixed 2 nm spectral bandwidth. The absorption amount was then expressed as mg of formaldehyde/kg of dry wood. **× ×**

**Figure 2.** Test apparatus for measurement of formaldehyde gases, type 2 for the flask method (1:500 mL bottle with a top made of polyethylene plastic; 2: hook to hang specimens; 3: rubber band to suspend specimens within the flask; 4: distilled water at the bottom of the flask): (**A**) the actual flask with a set of specimens; (**B**) linear drawing of the flask.

#### *2.4. Shear-Strength Test*

Specimens for the shear-strength test were prepared according to European standard specifications EN 314-1: 2004 (Figure 3). Specimens were tested using a universal test machine model STM-20, produced by Santam Engineering Design Co. (Semnan, Iran). Loading speed was 1 mm/min. Shear strength was calculated using Equation (1).

$$\text{S } \text{S} = \frac{F\_{\text{max}}}{A} \text{ (MPa)},\tag{1}$$

where *F*max is the maximum failing force, and *A* is the shear area in the specimen.

( )

,

max

**Figure 3.** Linear diagram of shear-strength test specimen according to European standard specifications EN 314-1: 2004.

#### *2.5. Delamination Test*

Delamination properties of the produced plywood were determined on the basis of voluntary standard specifications for plywood, provided by "The Hardwood Plywood and Veneer Associations (ANSI/HPVA HP-1, 2004). On the basis of the specifications, three specimens (50.8 mm × 127 mm) were cut from each panel. They passed three rounds of soaking/drying cycles. Each cycle comprised 4 h of soaking in water at 24 ± 2 ◦C, and then a drying period of 19 h with heating at 50 ± 1 ◦C. According to the specifications, a panel is acceptable for interior use if <5% of specimens delaminate after the first soaking/drying cycle. For exterior applications, delamination should not occur in >15% of the specimens. By definition, delamination is any continuous opening between two layers of plywood that is longer than 5.08 cm (or 2 in), deeper than 0.64 cm (or 0.25 in), and wider than 0.008 cm (or 0.003 in).

#### *2.6. Statistical Analysis*

One-way analysis of variance (ANOVA) was carried out in a completely randomized design and experiment with SAS software, version 9.2 (2010) at a 95% level of confidence. Duncan's multiple-range test was then performed to discern similar groupings among treatments for each property. Hierarchical cluster analysis was then performed using SPSS/18 (2010). Clusters included dendrograms by means of Ward's methods using squared Euclidean distance intervals.

#### **3. Results and Discussion**

#### *3.1. Preliminary Study*

In a preliminary study, 5%, 10%, 15%, and 20% UF resin was replaced with soy flour to gain a better estimation of the most effective combination to decrease formaldehyde emission. Results of the preliminary study demonstrated a clear decreasing trend in formaldehyde emission (FE) as soy flour content increased (Figure 4). This is in line with the data reported in the literature. Hosseini et al. [22] reported that the partial replacement of urea-formaldehyde adhesive with soy flour, particularly with a substitution rate of 15%, significantly reduced the formaldehyde emission. A similar observation was also made by Kawalerczyk et al. [23] who applied five types of flours (rye, hemp, coconut, rice, and pumpkin) as fillers with urea-formaldehyde resin in plywood manufacture. All soy-treated panels

showed a significant decrease in formaldehyde emission in comparison to control panels. The highest and lowest formaldehyde emissions were observed in the control panels (130.1 mg/kg) and panels with 20% soy flour (88.6 mg/kg), respectively. No significant difference was observed between formaldehyde emissions in panels produced with 10%, 15%, and 20% soy flour, although the FE trend was a decreasing one as soy flour content increased to 20%. Two factors potentially contributed to this behavior. The first one may have been the great dependence of formaldehyde emission on the resin content. A lower content of formaldehyde in the adhesive mixture causes a lower level of formaldehyde emission from the panel. In fact, a higher substitution of UF resin with soy flour results in a lower formaldehyde emission. Another factor is the reaction of free formaldehyde with amino groups present in the soy flour. Pereira et al. [37] found that the use of soy protein as a natural formaldehyde scavenger in wood particleboard production can contribute to a decrease in the formaldehyde content of particleboard panels, without significantly affecting the properties of the panels

**Figure 4.** Formaldehyde emission (mg/kg) in three-layer plywood panels produced with urea-formaldehyde resin and 5%, 10%, 15%, and 20% soy flour substitution of UF resin.

The shear strength of all panels produced in the preliminary phase of the study was also measured to find out the effects of the addition of soy flour on at least one mechanical property. Results showed that the shear strength values of all treatments were more than the standard limit of 1 MPa, even the control panels (1.55 MPa). The highest shear strength was observed in panels with 15% soy flour content (2.1 MPa); that is, soy flour resulted in a 34% increase in shear strength. The soy flour content of 5% did not have a significant effect on the shear strength of plywood panels (Figure 5). A further increase in soy flour content to 20% resulted in a decrease in shear strength (1.8 MPa), although it was still higher than that of the control specimens. Results of the delamination tests showed that all panels successfully passed the test (Table 2). This indicated that substitution of soy flour for UF resin (for soy flour contents lower than 20%) cannot significantly affect the delamination property of plywood panels. This finding is in accordance with a previously reported study, in which it was found that the partial replacement of UF adhesive with soy flour, particularly at substitution rates of 10% and 15%, did not significantly affect the shear strength of plywood under both dry and wet conditions [22].

Cluster analysis using the two properties of formaldehyde emissions and shear-strength values in the preliminary phase demonstrated similar clustering of the control panel with panels containing 5% soy flour (Figure 6). This indicated that 5% soy flour is too low to significantly influence the studied properties of plywood panels. The other three panels were distinctly clustered away from control panels. In order to benefit from the maximum shear strength, as well as a satisfactory decrease in

formaldehyde emission, the optimal soy flour contents of 10% and 15% were chosen for the next phase in which wollastonite was added to the UF resin along with soy flour.

**Figure 5.** Shear strength (MPa) in three-layer plywood panels produced with urea-formaldehyde resin and 5%, 10%, 15%, and 20% soy flour substitution of UF resin.


**Table 2.** Delamination test on plywood produced with different binders.

**<sup>1</sup>** No. of delaminated cases after the first round of the soaking/drying cycle. **<sup>2</sup>** No. of delaminated cases after the third round of the soaking/drying cycle. **<sup>3</sup>** P = testing passed, F = testing failed. **<sup>4</sup>** Urea-formaldehyde resin content. **<sup>5</sup>** Soy flour content.

**Figure 6.** Cluster analysis of five different three-layer plywood panels produced with urea-formaldehyde resin, and 5%, 10%, 15%, and 20% soy flour substitution for UF resin (Soy% = soy flour content).

#### *3.2. Main Phase of the Study*

In the second phase of the study, 5% and 10% wollastonite gel (W) was added to panels made with a mixture of UF resin and soy flour (only two optimal SF contents of 10% and 15% were tested in the main phase). Results illustrated that, apart from slight fluctuations which were attributed to the standard deviation among different specimens, the addition of wollastonite had no significant effect on formaldehyde emissions (Figure 7). However, both contents of wollastonite had a decreasing impact on shear strength (Figure 8), although the shear strength in W-added panels was still higher than that of control panels. The decrease in shear strength was attributed to the absorption of part of the resin by wollastonite particles. Results of the delamination tests revealed that all panels passed this test, indicating that the addition of SF or wollastonite did not have a significant impact on the delamination of plywood panels (Table 3).

**Figure 7.** Formaldehyde emission (mg/kg) in three-layer plywood panels (Soy% = soy flour content; W% = wollastonite content).

**Figure 8.** Shear strength (MPa) in three-layer plywood panels (Soy% = soy flour content; W% = wollastonite content).

Substitution of soy flour for UF resin at both levels of 10% and 15% significantly increased the water absorption (WA) and thickness swelling (TS) (Figures 9A and 10A after 2, 24 and 720 h and Figures 9B and 10B at various time intervals). The increases were significant at 2 h, 24 h, and the long-term immersion of 720 h. The increase was attributed to the water hydrophilicity of soy flour [38–41]. Addition of wollastonite (both W contents of 5% and 10%) to the resin mixture resulted in a significant decrease in water absorption, almost reaching the same value as in the control panels. This decrease was attributed to the reinforcing effect of wollastonite in the resin mixture. Similar reinforcing effects were previously reported to improve the shear strength of polyvinyl acetate resin [31] and the fire-retarding property of acrylic–latex paint [42]. With respect to thickness swelling, 5% wollastonite significantly decreased the TS to nearly the same level as in the control panels. However, TS values of panels containing 10% W were still nearly as high as the values of soy-treated panels. That is, a W content of 10% could not mitigate the negative effect of soy flour on thickness swelling. It was concluded that 10% wollastonite was too high, thereby absorbing UF resin rather than

acting as reinforcing filler to improve thickness swelling; therefore, a W content of 5% is recommended as an optimum.

**Table 3.** Delamination test on plywood produced with different binders and with the addition of wollastonite.


**<sup>1</sup>** No. of delaminated cases after the first round of the soaking/drying cycle. **<sup>2</sup>** No. of delaminated cases after the third round of the soaking/drying cycle. **<sup>3</sup>** P = testing passed, F = testing failed. **<sup>4</sup>** Urea-formaldehyde resin content. **<sup>5</sup>** Soy flour content. **<sup>6</sup>** Wollastonite content.

**Figure 9.** Water absorption (%) in three-layer plywood panels (Soy% = soy flour content; W% = wollastonite content; WA = water absorption).

The addition of both SF and W to UF resin resulted in a decrease in pH of the resin mixture (Table 4). Gel time was significantly increased as a result of the addition of both SF and W (Table 4); however, the gel times of all mixtures were below the hot-press time of 5 min. Therefore, it is unlikely that alterations in different properties measured in this study can be attributed to the difference in gel time. The addition of SF significantly increased viscosity, while W had a decreasing effect. The decreasing effect of W on viscosity was attributed to the water content of the W gel that was added to the resin mixture.



**<sup>1</sup>** UF = urea-formaldehyde resin; <sup>2</sup> SF = soy flour content; <sup>3</sup> W = wollastonite content

Cluster analysis as a function of the properties measured demonstrated that control panels were remotely clustered away from panels containing soy flour and wollastonite (Figure 11). This showed the significant effect of both soy flour and wollastonite on the overall properties of plywood panels. Panels containing either 5% or 10% wollastonite were also clustered differently from those containing only soy flour, indicating the significant impact of wollastonite. On the basis of the results of each property considered individually and altogether, it was concluded that panels containing 10% soy flour and 5% wollastonite are recommended to achieve the optimum decrease in carcinogenic formaldehyde emission, as well as the optimum physical and mechanical properties. With regard to the promising results of wollastonite as a reinforcing filler in different resins and coatings, further studies can be carried out to investigate the effects of the addition of wollastonite to different wood-based composite panels produced solely using bioresins, such as soy flour.

**Figure 11.** Cluster analysis of seven different three-layer plywood panels produced with urea-formaldehyde resin, and 10% and 15% soy flour substitution for UF resin, plus addition of wollastonite at 5% and 10% (Soy% = soy flour content; W% = wollastonite content).

#### **4. Conclusions**

Partial substitution of soy flour for urea-formaldehyde resin has the potential to decrease carcinogenic formaldehyde emission in plywood panels. Shear strength was also improved as soy flour content increased. However, this had a negative effect on the water absorption and thickness swelling of plywood panels. The addition of micron-sized wollastonite mitigated the undesirable increased hydrophilicity in panels caused by soy flour. It was concluded that 10% soy flour and 5% wollastonite provide the lowest formaldehyde emission and the most optimum physical and mechanical properties.

**Author Contributions:** Methodology, H.R.T. and S.G.; validation, H.R.T., S.G., M.G., and S.B.H.; investigation, H.R.T. and S.B.H.; writing—original draft preparation, H.R.T., and A.N.P.; writing—review and editing, H.R.T., S.B.H., and A.N.P.; visualization, H.R.T. and S.B.H.; supervision, H.R.T. and A.N.P. All authors read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The first author appreciates the constant scientific support of Jack Norton (retired, Horticulture and Forestry Science, Queensland Department of Agriculture and Fisheries, Australia), as well as Alexander von Humboldt Stiftung (Bonn, Germany).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Alien Wood Species as a Resource for Wood-Plastic Composites**

**Sergej Medved, Daša Krapež Tomec , Angela Balzano and Maks Merela \***

Department of Wood Science and Technology, Biotechnical Faculty, University of Ljubljana, Jamnikarjeva 101, 1000 Ljubljana, Slovenia; sergej.medved@bf.uni-lj.si (S.M.); dasa.krapez.tomec@bf.uni-lj.si (D.K.T.); angela.balzano@bf.uni-lj.si (A.B.)

**\*** Correspondence: maks.merela@bf.uni-lj.si

**Abstract:** Since invasive alien species are one of the main causes of biodiversity loss in the region and thus of changes in ecosystem services, it is important to find the best possible solution for their removal from nature and the best practice for their usability. The aim of the study was to investigate their properties as components of wood-plastic composites and to investigate the properties of the wood-plastic composites produced. The overall objective was to test the potential of available alien plant species as raw material for the manufacture of products. This would contribute to sustainability and give them a better chance of ending their life cycle. One of the possible solutions on a large scale is to use alien wood species for the production of wood plastic composites (WPC). Five invasive alien hardwood species have been used in combination with polyethylene powder (PE) and maleic anhydride grafted polyethylene (MAPE) to produce various flat pressed WPC boards. Microstructural analyses (confocal laser scanning microscopy and scanning electron microscopy) and mechanical tests (flexural strength, tensile strength) were performed. Furthermore, measurements of density, thickness swelling, water absorption and dimensional stability during heating and cooling were carried out. Comparisons were made between the properties of six WPC boards (five alien wood species and mixed boards). The results showed that the differences between different invasive alien wood species were less obvious in mechanical properties, while the differences in sorption properties and dimensional stability were more significant. The analyses of the WPC structure showed a good penetration of the polymer into the lumens of the wood cells and a fine internal structure without voids. These are crucial conditions to obtain a good, mechanically strong and water-resistant material.

**Keywords:** alien plants; wood plastic composite; flexural strength; tensile strength; swelling; dimension stability; scanning electron microscopy

#### **1. Introduction**

The properties of wood-based composites are determined by the components used for their production. This is also demonstrated in the case of wood plastic composites (WPC), where wood can act as a reinforcement or as a filler and in some cases both. Wood and the derived components are an important factor influencing the properties of wood-based panels [1–5]. WPC is basically composed of two main components, namely plastic or polymer and wood, resulting in a material which combines the best properties of both components. Although the wood constituents in WPC are small (usually size class between 0.1 mm and 1.0 mm), and (according to [6,7]) the wood species related differences should be smaller, several authors [8–13] have shown that the wood constituents (in terms of species and size of constituents) influence the properties of WPC. The influence of the wood species used for WPC depends on the size of constituent obtained during breakdown process (particularly in terms of slenderness ratio), its affinity towards polymeric compound and strength of bond between wooden constituent and polymeric matrices. Shebani et al. [14] determined that chemical composition of wood also influences the properties of WPC; namely, they proved that a higher cellulose and lignin content results in better mechanical

**Citation:** Medved, S.; Tomec, D.K.; Balzano, A.; Merela, M. Alien Wood Species as a Resource for Wood-Plastic Composites. *Appl. Sci.* **2021**, *11*, 44. https://dx.doi.org/10.3390/app 11010044

Received: 27 November 2020 Accepted: 18 December 2020 Published: 23 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/).

properties but also in lower moisture resistance when the cellulose content is high. One of the disadvantages of using wood in WPC is the reaction of the wood to UV radiation. The effect of UV radiation (e.g., sunlight) on the wood surface leads to photochemical degradation of wood and thus of WPC. Colour changes (darkening) also occur during the production of WPC. Exposure to elevated temperatures during the pressing process leads to the evaporation of extractives, which darken the wood surface and thus also WPC [10,15].

Based on data from the existing Flora of Slovenia (CCFF) Database, the species *Robinia pseudoacacia* (black locust) is the invasive alien plant species with the potentially most negative impact on biodiversity. Tree of heaven (*Ailanthus altissima*) and boxelder maple (*Acer negundo* L.) can also be classified as invasive alien plants with a high negative impact on biodiversity. It is, therefore, undoubtedly useful to raise awareness of the impact of invasive alien plant species on our environment and to look for the most versatile applications, including their use as WPC components.

Our main objective is to test the suitability of the most widespread invasive alien hardwood species present in Slovenia for the production of WPC and to encourage their removal from native natural ecosystems by transforming them into a source of raw materials that could be processed into useful products. The mechanical properties of these wood species are not well known, and further information about them may encourage the use of these woods in the most appropriate way in new products. Wood anatomical analyses and machining tests have already been carried out on the same wood species, which overall showed good degree of machinability [16,17]. In light of the results obtained so far, we believe that the selected wood species may be suitable for the production of WPC due to its anatomical structure and mechanical properties. To verify their suitability for this application, we carried out classical mechanical tests, such as flexural and tensile strength tests. In addition, we performed microstructural analyses of the surface and internal structure of WPC boards using scanning electron microscopy (SEM).

SEM is a powerful tool for examining the surface and structure of wood. The application of SEM in wood science is well described in the literature [18–22]. Recently SEM has been successfully used to study the morphology and surface evaluation of WPC [13,23]. SEM has been used to evaluate the adhesion between wood and polymer matrix and detect the occurrence of fibre pull outs and voids within the composite [24,25]. It was shown that the evaluation of SEM is consistent with the sorption behaviour and can clearly explain the mechanical properties of WPC. It was observed that an intact composite surface corresponds to a lower moisture transport rate within the matrix [26]. Therefore, an intact and homogeneous material with stronger adhesion of its two components, namely wood and polymer matrix, results in a material with higher mechanical properties. Given the reported advantage of the microscopy method, we used Confocal laser scanning microscopy (CLSM) and Scanning electron microscopy (SEM) for the surface analysis of WPC boards, to evaluate their homogeneity, the quality of adhesion at the interface between the wood fibres and the polymer matrix, and the possible occurrence of voids that could reduce their mechanical properties. We used the results to discuss the mechanical properties as well as the sorption properties of WPC boards considering the observed microscopic features.

#### **2. Materials and Methods**

#### *2.1. WPC Boards Preparation*

We prepared WPC boards using some of the most widespread invasive alien hardwood species present in Slovenia, namely: boxelder maple (*Acer negundo*), horse chestnut (*Aesculus hippocastanum*), tree of heaven (*Ailanthus altissima*), black locust (*Robinia pseudoacacia*) and honey locust (*Gleditsia triacanthos*). The boards were produced using wood of the aforementioned species and polyethylene (PE) powder DowlexTM 2631.10UE obtained from local company ROTO-Pavlinjek d.o.o. (MURSKA SOBOTA, Slovenia). The physical properties of the powder used are shown in Table 1.


**Table 1.** Physical properties of polyethylene powder DowlexTM 2631.10UE (values were determined by the supplier).

−

Maleic anhydride grafted polyethylene (MAPE) was used as a coupling agent (donated by Graft Polymere d.o.o., Ljubljana, Slovenia). MAPE was added to increase the affinity and adhesion between wooden constituents and polymeric matrices. MAPE acts as so-called coupling agent or compatibilizing agent.

The polyethylene content was 46.5% and MAPE 3.5%, while 50% of the total mass consisted of wood particles obtained from 5 alien wood species with different densities [17]:


The 50% wood content was selected based on a report by Leu et al., 2012 [27], which showed that the mechanical properties of WPC increased by up to 50%, while a higher share led to a decrease in mechanical properties.

A two-step decomposition process was used to break down wood into particles (Figure 1).

**Figure 1.** Wood break down process.

The breakdown of solid wood into chips was carried out in a Prodeco M-0 chipper, which has an output screen with openings of 25 mm in diameter. The ring chipper used for production of particles was a Condux CSK 350/N1 ring chipper (the gap between the blade and beating bar was 1.25 mm). After chipping particles were analysed by sieving, whereby 100 g of particles were placed on the top sieve. After 10 min of sieving, the residues on each sieve were weighted. The particles used for the experiment are shown in Figure 2.

As the moisture content of the particles was higher than required for WPC production, the particles were dried at 80 ◦C for 24 h to achieve a moisture content below 4% (the actual moisture content for board production was between 0.9% and 2.3%). After drying, the par-

ticles were mixed by hand with PE powder and MAPE. The mass ratio was 50:46.5:3.5 (wood:PE:MAPE).

**Figure 2.** Wood particles used for research; (**a**) black locust, (**b**) boxelder maple, (**c**) honey locust, (**d**) horse chestnut, (**e**) tree of heaven.

The prepared mixture was hand formed into a frame measuring 300 <sup>×</sup> 300 mm<sup>2</sup> , which was placed on a steel plate. The target thickness was 4 mm, the target density 0.9 g·cm−<sup>3</sup> .

− Wood-PE mat was flat pressed at 180 ◦C for 10 min at a specific pressure of 3 MPa. After 10 min the boards were transferred to the cold press. The specific pressure during cold pressing was the same as during hot pressing (3 MPa), while the pressing temperature was set at 25 ◦C (equal to room temperature). The cooling process also hardened the PE. The process for preparing the WPC is shown in Figure 3.

**Figure 3.** Schematic layout of wood plastic composites (WPC) preparation.

Six sets of WPC boards were prepared from different wood species as shown in Table 2:

−

−


**Table 2.** WPC board types regarding wood species used.

In the mixture, 20% of each wood species was used. After climatization period, boards were cut for testing:


#### *2.2. Physical and Mechanical Properties Testing*

Flexural strength was determined by a three-point bending test on the testing machine Zwick Roell Z005 testing machine. Since particles were evenly distributed over the width and length of board and no difference in fibre orientation was expected, only one direction was tested. The span distance was 64 mm, while the loading speed was set to 2 mm·min−<sup>1</sup> . Maximum force, deformation at maximum force, flexural strength and modulus of elasticity were determined.

The tensile strength was determined on Zwick Roell Z005 testing machine, also in one direction only. The loading speed was set to 5 mm·min−<sup>1</sup> . Maximum force, deformation at maximum force and tensile strength were determined.

Thickness swelling and water absorption were determined by immersion of samples in water. The immersion time was 2 and 24 h. Thickness swelling (TS) in % and water absorption (WA) in % were calculated by Equations (1) and (2):

$$TS\_y = \frac{t\_2 - t\_1}{t\_1} \times 100\tag{1}$$

$$
\Delta W\_y = \frac{m\_2 - m\_1}{m\_1} \times 100\tag{2}
$$

where *y* represents immersion time, *t* sample thickness in mm, m mass of samples in g, while 1 denotes the thickness or mass before and 2 the thickness or mass after 2 h or 24 h of immersion.

The dimensional stability of samples was determined by exposing one set of 5 samples to a temperature of −25 ◦C and one set of 5 samples to a temperature of +65 ◦C. The exposure time was 60 ± 1 min. The dimensional stability (*δx*) in % was calculated by Equation (3)

$$
\delta\_{\mathbf{x}} = \frac{\mathbf{x}\_2 - \mathbf{x}\_1}{\mathbf{x}\_1} \times 100 \tag{3}
$$

where *x* represents length or thickness respectively, while 1 denotes the dimension before and 2 the dimension after 60 ± 1 min exposure. All results were evaluated using Statistica software by ANOVA and LSD test at α = 0.05.

#### *2.3. WPC Structural Analyses*

To evaluate WPC surface and internal structure, sub-samples of boards were prepared and observed using a Confocal Laser Scanning Microscope (CSLM) and a Scanning Electron Microscope (SEM). For the structural analyses we used WPC boards made of a

mixed material (different wood species). Before the observation, the sub-samples were cut on their cross-section surface with a blade on a sliding microtome (Leica SM2000, Nussloch, Germany) to obtain a flat and smooth surface and then dried at room temperature (T = 22 ◦C and RH = 65%) [16]. To obtain a panoramic view of the sub-sample and to inspect its entire surface, it was placed on the stage of Confocal Laser Scanning Microscope (CLSM) Olympus LEXT OLS5000 (Olympus Corporation Tokyo 163-0914, Tokyo, Japan) and observed with the optical system using the MPLFLN10xLEXT objective (numerical aperture 0.3, working distance 10.4 mm). Images of the entire surface area were obtained by combining several images at different focus positions, which were recorded in real time using the stitching function by moving the stage. SEM was used to investigate the quality of adhesion at the interface between wood fibres and the polymer matrix and to detect possible voids. Before SEM observations, samples were mounted on stubs with a conductive carbon adhesive tape and coated with an Au/Pd sputter coater (Q150R ES Coating System; Quorum technologies, Laughton, UK) for 30 s with a constant current of 20 mA. The SEM micrographs were then recorded in a high vacuum with 5 kV voltage and with a large field detector (LFD) in a FEI Quanta 250 SEM microscope (FEI Company, Hillsboro, OR, USA) at 9.3 mm working distance and at 100×, 250×, 500× and 2500× magnification.

#### *2.4. Pilot Production of 3D Composites Based on the Proposed Methodology*

Wood residues, which arise from the primary processing of wood and the production of wood products, were firstly chipped in a mill and secondly in a knife ring chipper (as presented in Figure 1). Subsequently, the obtained particles were additionally ground with the Retsch SM2000 rotary wood mill (Retsch, Haan, Germany) with a 1 mm sieve.

The particles were then dried at 80 ◦C for 24 h to achieve a moisture content of less than 4%. Polyethylene (PE) DowlexTM powder, maleic anhydride grafted polyethylene (MAPE) and wood particles obtained from 5 different invasive alien wood species were used in a ratio of 46.5:3.5:50 (PE:MAPE:wood).The mixture was formed by hand into a 3D mould, which was primarily sprayed with a non-stick agent (Silicone H1 spray, Panolin, Madetswil, Switzerland) and then pressed in a hot press at 180 ◦C and a specific pressure of 3 MPa for 10 min. The mould was then transferred to the cold press with the same pressing parameters, which differ only in temperature, namely 25 ◦C.

The 3D WPC product was then removed from the mould and, where necessary, edge milling and sealing was carried out.

#### **3. Results and Discussion**

Most of particles (65–70%) used in experiment was size class 1.5 and lower, as classified by screening (particles that fell through sieve with opening 2.0 mm), while minority of particles (30–35%) was size class 2 mm and higher (Table 3).


**Table 3.** WPC board types regarding wood species used.

Although particles were prepared under same conditions, there are differences between them related to the particle size class (share of residue on the sieve).

#### *3.1. Physical and Mechanical Properties of WPC Boards*

The properties of WPC depend on the polymer type used and the type of wood species used for its production. Since polymer was the same the differences between WPC bords (Figure 4) are caused by wood species used through their structure (chemical and anatomical), generated particles, their compressibility, interaction with polymer and through their mechanical properties (Tables 4 and 5).

**Figure 4.** Produced WPC; (**a**) boxelder maple, (**b**) horse chestnut, (**c**) tree of heaven, (**d**) black locust, (**e**) honey locust, (**f**) mixture. Samples are 50 <sup>×</sup> 50 mm<sup>2</sup> .

**Table 4.** Physical and strength properties of flat pressed WPC (letters in bracket denote same homogeneous group determined by LSD test at α = 0.05).


**Table 5.** Sorption properties and dimensional stability of flat pressed WPC (letters in bracket denote same homogeneous group determined by LSD test at α = 0.05).


α

**−**

α

The PE matrix (or the matrix of WPC in general) is responsible for the load transfer between constituents and moisture resistance, while wood is responsible for density, strength and stiffness. The result (properties of WPC) should be the combination of the best properties of the components. The impact of the wood species can already be seen in the WPC density (Table 4 and Figure 5), where the densities of WPC boards made of higher density wood species differ from those of lower density and mixture. With regard to the density of the wood species, we can divide the material into two different categories, namely wood species below 0.6 g·cm−<sup>3</sup> , and those with higher density. From this perspective, we can observe an interesting relationship between the density of the wood itself and the density of the WPC (Figure 5). −

**δ δ − δ δ −**

− − − − − − − − − − − − − − −

− − −

**Figure 5.** Density of WPC with respect to wood species used (A—boxelder maple; B—horse chestnut; C—tree of heaven; D—black locust; E—honey locust; F—mixture).

The highest increase in density was at use of the horse chestnut (compaction ratio (Ratio between WPC density and density of wood species) 1.88), while the lowest was in black locust (compaction ratio 1.23), which was also expected. Wood species with low density are indeed more compressible at the same condition as those with higher density. Furthermore, we can observe from Figure 5 that an increase in density results in a decrease in WPC. Again, we can notice two different set namely behaviour of wood species with a density below and those above 0.6 g·cm−<sup>3</sup> . The decrease in density is more pronounced for boards made from wood species with low density, while differences are smaller for boards made from wood species with density above 0.7 g·cm−<sup>3</sup> . Such behaviour could be related to compressibility of wood or to penetration of polymer matrix into cell lumens. Tangential diameter of vessel lumina are: 50–100 µm in *Acer negundo*, ≤50–100 µm in *Aesculus hippocastanum*, ≥200 µm in *Ailanthus altissima*, 100–≥200 µm in *Robinia pseudoacacia* (with common tyloses) and 100–≥200 µm in *Gleditsia triacanthos* [32]. The polymer compound penetrated easily into larger lumens at more dense wood species, while at wood species with lower density the compression of cell wall occurred prior to mobilization of polymeric compound into cell lumens. For low density wood species, the increase in density is due to the compression of the cell walls, while for high density wood species the penetration of PE in lumens lead to a higher density (although the compression ratio is lower).

The differences between alien wood species are less obvious in terms of flexural (Figure 6; *p* value 0.40) and tensile strength (Figure 7; *p* value 0.06), while the differences in modulus of elasticity (Figure 8; *p* value 0.00), sorption properties (Figures 9 and 10; *p* value 0.01 respectively 0.00) and dimensional stability (Table 4) are more significant (at α = 0.05).

−

≥ ≥

≥

−

≤

α

**Figure 6.** Flexural strength of WPC with respect to wood species used (A—boxelder maple; B—horse chestnut; C—tree of heaven; D—black locust; E—honey locust; F—mixture).

**Figure 7.** Tensile strength of WPC with respect to wood species used (A—boxelder maple; B—horse chestnut; C—tree of heaven; D—black locust; E—honey locust; F—mixture).

**Figure 8.** Modulus of elasticity of WPC with respect to wood species used (A—boxelder maple; B—horse chestnut; C—tree of heaven; D—black locust; E—honey locust; F—mixture).

**Figure 9.** Thickness swelling and water absorption after 2 h immersion (A—boxelder maple; B—horse chestnut; C—tree of heaven; D—black locust; E—honey locust; F—mixture).

**Figure 10.** Thickness swelling and water absorption after 24 h immersion (A—boxelder maple; B—horse chestnut; C—tree of heaven; D—black locust; E—honey locust; F—mixture).

The WPC strengths (flexural and tensile) are related to the combined effect of particle size (Figure 2, Table 4) and WPC density. Medved et al. [33] determined size related differences between some alien wood species. They analysed particles of some alien wood species. Black locust and staghorn sumac gave the longest particles, while the shortest were determined at honey locust and tree of heaven. The authors also determined differences in aspect ratio (ratio between particle length and width), tree of heaven gave the particles with the lowest aspect ratio, while black locust the highest. In wood species with lower density the densification (higher compaction ratio) enabled optimal strength values, whereas for WPC made from higher density wood species (D and E series), adequate particle morphology and their mechanical properties are to be considered. In the case of flexural strength, two values stand out, namely in the case of WPC made of black locust (D series) and mixture (F series), one having the highest value and the other the lowest. The high value at black locust (D series) is related to its density and high strength properties, while the low flexural strength of the WPC board made of mixture is the consequence of its low density. An important aspect of WPC strength is related to the size of constituent, its embedding and interaction with the polymer matrix. In order to achieve adequate strength, the fibrous material (in our case wood particles), must be long enough to resist the forces applied to them, especially the shear and tensile forces generated when the

fibrous elements are pulled out of the matrix. According to Callister [34], the most likely occurrence of failure is the end of fibrous elements, where the shear stresses are highest and the tensile stresses are lowest. The load is transferred from the matrix through the particle ends through shear, which gradually "moved" to tensile, which was more carried by the particle and less by the matrix. In such a loading behaviour, the aspect ratio of the particles is important, i.e., when the aspect ratio is low, the load transfer overlaps at the ends, so that the strain gradient in the particles does not reach the strain gradient in the PE matrix. [35,36] When the particles are long enough, their ability to withstand the load is much higher, and according to the flexural strength results, black locust particles have reached and exceeded this critical particle length. Although modulus of elasticity and tensile strength should follow the same pattern (related to particle dimensions and density), the results in our experiment do not support this. The highest tensile strength was determined at WPC made of mixture (F series), and the highest modulus of elasticity was found for WPC made of horse chestnut (B series). The lowest modulus of elasticity was determined for WPC made of mixture (F series), and the lowest tensile strength for boxelder maple (A Series).

The differences in the strength properties of WPC could be related to the particle morphology, its densification rate as well as to its interaction with the polymer matrix. A possible reason for the differences could also be the presence of micro- and macro voids in the particles formed during disintegration and drying. Such micro- and macro voids lead to a strength reduction due to a less efficient load transfer from the matrix to the particles. The comparison of the strength properties of WPC made from alien wood species with WPC made from spruce (*Picea abies*) shows similar values. The properties of WPC made from spruce are presented in Table 6. WPC board made of spruce wood was in our laboratory made by same conditions and process as compared boards made of invasive species.


**Table 6.** Comparison of WPC board properties from spruce wood with properties of WPC from alien wood species.

The influence of the type of wood used for WPC was also determined at dimensional stability and moisture resistance.

Although the particles are embedded into PE matrix, water can penetrate the particles to cause the thickness change. However, the degree of change (thickness swelling and water absorption) could be related to the sorption properties of the wood. We assume that the differences are more related to the interaction between wood and PE matrix. In the case of a good interaction, the PE matrix efficiently embedded the wood particles and penetrated into the cell wall lumens, sealing them and thus making them inaccessible to water penetration into the lumen, and the effect (thickness swelling and water uptake) was lower. The amount of water absorbed by wood depends strongly on the number of free hydroxyl groups to which water can attach, and if the bond between PE matrix and wood is compact, then these hydroxyl groups are occupied by another component and therefore unavailable to water. The differences in thickness swelling could also be related to the composite density. Lower swelling was observed for WPC with densities between 0.92 g·cm−<sup>3</sup> and 0.96 g·cm−<sup>3</sup> .

The sensitivity of the particles to moisture, although embedded in the PE matrix, was also determined by dimensional stability test (Table 5). When exposed to a temperature of +65 ◦C, the size of the WPC decreases in length and thickness (compared to the value

before exposure). We hypothesise that this could be related to the shrinkage of the particles, while exposure to lower temperatures (−25 ◦C) causes an expansion in length that could be related to the expansion of water molecules in the cell wall, in the lumen of the particles, between the particles, and between the particles and the PE matrix. Water is indeed an exception when exposed to low temperatures, namely when water freezes, it can expand by about 9% [37].

#### *3.2. Surface and Internal Structure of WPC Boards*

According to the methodology presented above, we investigated the surface and internal structure of WPC boards made of wood species mixture (F series). First Scanning Electron Microscopy (SEM) observations were performed on the unflattened surface of WPC boards. In this case, we could only observe flat homogeneous surface and we could not recognize any wood structure, especially at lower magnification (100×) (Figure 11a). Rarely, in some areas, we could observe some voids on the surface. A detailed analysis (1000× magnification) revealed some wood anatomical structure which were, however, difficult to recognize and could not enable the identification of wood species (Figure 11b).

**Figure 11.** Scanning Electron Microscopy observation on the unflattened surface of WPC boards; (**a**) homogeneous surface with no recognizable wood structure, (**b**) higher magnification image revealing some wood structures.

The main objective of the microscopic analysis of the WPC structure was to observe the interaction and adhesion between the polymer matrix and the small wood particles embedded in it. To enhance the observation of the anatomical structures of the wood, WPC cross-section were pre-treated. A first overview of a WPC pre-treated cross-section was made on the region of interest (ROI) by a Confocal Laser Microscope (Figure 12). On Figure 12 we can clearly observe wood particles. We could recognize the polymer matrix between the wood particles and within the wood pores. In the WPC cross-section we could not detect any voids. To observe in detail and with high resolution the interaction between wood particles and polymer we used SEM. Various regions of interest (ROI), which we observed at SEM, are marked A1 to A4 in Figure 12.

In Figure 13 the internal WPC board structure at 100× and 250× magnification is shown (ROI marked on Figure 11). In this case, the structure of wood is clearly identifiable, so we can recognize the arrangement of the vessels as well as the wood fibres and tracheids. According to previous wood structure analysis of IAPS [16,17] we can identify boxelder maple (*Acer negundo*) cross section. This proved that with our SEM methodology it is possible to identify the wood species used for WPC production although the wood particles are small.

A larger magnification (500× and 2500×) in Figure 14 revealed interaction between wood particles and polymer matrix. The embedding of wood components in a polymer matrix, as well as lumens filled with polymer, is clearly visible. In Figure 14b is shown a detailed cross-section structure, in which we can see that all fibres, as well as ray parenchyma

cells, are filled with polymer. The absence of considerable voids between wood fibre and matrix indicates good compatibility and good interfacial adhesion.

**Figure 12.** Confocal Scanning Laser image of a cross section WPC board. A1 to A4 are ROI analysed by Scanning Electron Microscopy (scale bar is 500 µm).

**Figure 13.** Scanning Electron Microscopy image of a cross section WPC board; (**a**) A1 analysed ROI, (**b**) A2 analysed ROI.

The detailed structural observation of WPC showed that the WPC production process used (flat pressing) produced a fine and filled structure without large any major voids or fibre pull out, and with good interfacial adhesion, so that the moisture resistance, mechanical properties and thermal stability were relatively high.

The experience gained in the production of flat pressed WPC was used to carry out a pilot production of 3D shaped composites. Following the described procedure, we produced several different 3D-shaped WPC products made from invasive alien woody plants (Figure 15).

PE

PE

(**a**) (**b**)

**Figure 14.** Scanning Electron Microscopy image of a cross section WPC board. (**a**) A3 analysed ROI, (**b**) A4 analysed ROI.

**Figure 15.** Some examples of 3-dimensional shaped WPC products made fr **Figure 15.** Some examples of 3-dimensional shaped WPC products made from invasive alien woody plants.

Based on the results of this research and the newly gained experience, we will further develop the production of wood-based composites from invasive species, analysing the impact of particle morphology differences in order to optimise (increase) the proportion of wood in WPC.

Future studies on replacing polyethylene with polylactic acid (PLA)—A biodegradable, renewable material derived from crops such as corn and sugarcane, would be of great interest. PLA is one of the fastest growing bioplastics in the bio-composites industry due to its good properties such as renewability, biodegradability, biocompatibility, ease of processing and high modulus [38].

#### **4. Conclusions**

The present study shows that the differences between the mechanical properties of WPC were less pronounced in all the selected invasive alien wood species studied, while the differences in sorption properties and dimensional stability were more significant. We conclude that good adhesion and complete embedding of the polymer material in the wood cells is crucial to obtain a good, solid, mechanically strong and water resistant WPC material. Invasive alien plant species proved to have a high potential for the production of WPCs. Taking into account the economic indicators, it is currently difficult to demonstrate a high added value of the developed products, but there is the potential to do so in the future. In the processing of wood residues from invasive alien plant species, it is the reuse of harmful invasive alien plants that brings a particular added value to our products, as it contributes (in)directly to the care of the environment and the conservation of biodiversity. That is the greatest contribution of this work.

**Author Contributions:** Conceptualization, S.M. and M.M.; methodology, S.M., D.K.T., A.B. and M.M.; validation, S.M. and M.M.; experimental analyses, D.K.T. and A.B., formal analysis, S.M. and M.M.; writing—original draft preparation, S.M. and M.M.; writing—review and editing, S.M., D.K.T., A.B. and M.M.; visualization, S.M. and M.M.; supervision, S.M. and M.M.; project administration, M.M.; funding acquisition, M.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research was supported by APPLAUSE (UIA02-228) project, co-financed by the European Regional Development Fund through the Urban Innovative Actions Initiative (www. ljubljana.si/en/applause/), and additionally supported by the Program P4-0015, co-financed by the Slovenian Research Agency.

**Acknowledgments:** The authors wish to thank Jože Planinšiˇc, Luka Krže and Denis Plavˇcak (production of 3D WPC) for their immense help with sample preparation.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Article*
