Industrial byproducts as adhesive allies: Unraveling the role of proteins and isocyanates in polyurethane wood bonding

Mary, A., Blanchet, P., Pepin, S., Hermann, A., Charron, S., and Landry, V. (2024). “Industrial byproducts as adhesive allies: Unraveling the role of proteins and isocyanates in polyurethane wood bonding,” BioResources 19(2), 3520-3542.

Abstract

Wooden structures are becoming increasingly popular in the construction world. However, these structures often rely on synthetic adhesives, raising concerns about the environmental risks associated with their chemical composition. In response to these concerns, this study aims to explore sustainable alternatives, particularly focusing on polyurethane adhesives that incorporate proteins from industrial byproducts. The investigation involved three protein sources: soybean meal, shrimp shells, and skim milk, modified under mild alkaline conditions to obtain protein concentrates. These concentrates were then incorporated into the adhesives at varying protein contents: 5%, 10%, and 15%. Additionally, two isocyanate systems were examined, one being petrochemical-based and the other a partially bio-based blend. Chemical, thermal, optical, and mechanical characterizations were conducted to evaluate the adhesive performance. This study demonstrates that the adhesives’ thermal properties remain unaffected by both the protein content and the isocyanate system. However, these factors influence the adhesive penetration into the wood substrate. Ultimately, the results suggest that higher protein content offers superior retention of mechanical strength in adhesives compared to the petrochemical reference when subjected to humid conditions. Overall, this research demonstrates the potential of proteins from industrial byproducts as sustainable adhesive allies, providing valuable insights into their interactions with different isocyanates.

Download PDF

Full Article

Industrial Byproducts as Adhesive Allies: Unraveling the Role of Proteins and Isocyanates in Polyurethane Wood Bonding

Alex Mary,^a,b Pierre Blanchet,^a,b Simon Pepin,^a,b Aurélien Hermann,^b Stéphane Charron,^c and Véronic Landry ^a,b,*

DOI: 10.15376/biores.19.2.3520-3542

Keywords: Wood; Engineered wood products; Polyurethane adhesive; Isocyanate; Byproducts; Proteins

Contact information: a: NSERC – Industrial Chair on Eco-Responsible Wood Construction (CIRCERB), Université Laval, Québec, QC G1V 0A6, Canada; b: Department of Wood and Forest Sciences, Université Laval, Québec, QC G1V 0A6, Canada; c: Buildwise (ex-BBRI), 1342, Limelette, Belgium;

*Corresponding author: veronic.landry@sbf.ulaval.ca

GRAPHICAL ABSTRACT

INTRODUCTION

Wood structures are experiencing a surge in popularity due to their versatility and sustainability potential (Raydan et al. 2021). However, the conventional reliance on petroleum-based adhesives for structural wood products poses significant environmental challenges. These petroleum resources deplete finite reserves and contribute to harmful environmental impacts, exacerbating concerns over climate change and pollution. In response to these pressing environmental issues, the primary motivation of this research is to develop more sustainable and environmentally responsible adhesive solutions for wood bonding. In this context, biopolymers such as proteins, tannins, lignin, starch, and oils, have emerged as promising candidates for sustainable adhesive formulations (Pizzi and Mittal 2005; Calle et al. 2016; Gui et al. 2016; Vnučec et al. 2017; Luo et al. 2020; Arias et al. 2022).

Proteins, which can be of plant or animal origin, have attracted increased interest in the potential improvement of industrial wood adhesives. They are abundant biomass resources with low toxicity and a reduced environmental footprint and can improve adhesive adhesion to wood (Yang et al. 2006; Raydan et al. 2021). Adhesives derived from plant-based proteins, including wheat gluten, soybean meal, and cottonseed, exhibited commendable adhesive strength but poor water resistance (Raydan et al. 2021). In contrast, animal-origin adhesives, such as those derived from blood, caseins (milk proteins), keratin, and collagen, demonstrated improved water resistance compared to their vegetable counterparts (Raydan et al. 2021). However, they come with drawbacks, such as a non-homogeneous composition with the formation of agglomerates that may affect the integrity of the glue line (Raydan et al. 2021).

Studies, including this one, are underway to develop protein-based adhesives that have properties comparable to their petrochemical counterparts while being more environmentally friendly. To do so, polyurethane adhesive was selected for this study as a reference adhesive for several reasons. It allows the assembly of a wide range of materials, can form hydrogen and covalent bonds with the wood substrate, does not require formaldehyde, and its small molecules allow the penetration of porous substrates (Pizzi and Mittal 2005). Furthermore, two-component polyurethane adhesives, consisting of a combination of isocyanate and polyol components, are flexible and are recommended for assemblies subject to high mechanical and thermal stress. Proteins were used to substitute the OH groups of the polyols with NH groups present in the proteins. This substitution leads to the formation of urea (C-N) bonds (Kumar and Pizzi 2019). As concerns regarding petroleum resource depletion and environmental protection continue to grow, using biological byproducts generated by industries offers a promising sustainable approach (Bai et al. 2020). Thus, three protein sources were used: soybean meal, shrimp shells, and skim milk powder, and incorporated for polyol replacement into the adhesives at varying protein contents. These three sources are industrial co-products. Soybean meal and shrimp shells are non-recyclable, while skim milk powder is in excess in Canada, due to higher demand for cream and butter, with nearly 19 tons of product exported in 2022 according to the Government of Canada. This incorporation follows on from and is described in a previous study (Mary et al. 2024). Additionally, the study examines two different isocyanate systems: one employing a petrochemical-based isocyanate and the other utilizing a partially bio-based blend comprising 75% petrochemical isocyanate and 25% partially bio-based isocyanate. This research aims to provide insights into the influence of proteins and isocyanates on adhesive performance through comprehensive chemical and mechanical characterization.

EXPERIMENTAL

Materials

The raw materials used as protein sources were non-recyclable or excess industrial byproducts from local resources. Soybean meal in pellet form was obtained from Sollio Agriculture (Lévis, Canada). Shrimp shell flour was provided by Les Pêcheries Marinard ltée (Gaspé, Canada). Skim milk powder was provided by Agropur (Longueuil, Canada). Polymeric methylene diphenyl diisocyanate (Mondur MR-5) (NCO wt% = 32.5, viscosity = 129 mPa.s at 20 °C, typical functionality = 2.4) and a polypropylene oxide-based triol (Multranol 8175) (acid value = 350 to 390 mg KOH/g sample, molecular weight = 450 Da, viscosity = 232 to 412 mPa.s at 25 °C) were purchased from Covestro (Pittsburgh, PA, USA), provided by EMCO-Inortech (Terrebonne, Canada). Tolonate ™ X FLO 100, an aliphatic isocyanate polymer (NCO wt% = 12.3 viscosity = 140 ± 80 mPa.s at 25 °C, functionality = 2) from Vencorex (Saint-Priest, France), was provided by EMCO-Inortech (Terrebonne, Canada). Mondur MR-5 is a petrochemical isocyanate, noted PI, and Tolonate ™ X FLO 100, a partially bio-based and solvent-free isocyanate, noted as BI. All chemicals were used as received. The wood species used in this study was spruce wood (Picea abies, Karst) obtained from Buildwise (Limelette, Belgique).

Proteins Extraction

Protein concentrates were prepared for soybean meal and shrimp shells by alkaline extraction. The hydrolysis, washing, and centrifugation steps were detailed in a previous study (Mary et al. 2024). Based on precipitation and drying, a different protocol was used for skim milk powder (Husnaeni et al. 2019). The protein concentrates were then ground and passed through a 200 µm sieve. A previous study analyzed the properties of these protein concentrates, which are listed in Table 1. Protein concentrates derived from soybean meal, shrimp shells flour, and skim milk powder are designated B, S, and M, respectively.

Glycinin, β-conglycinin, and actin are globular proteins, while caseins behave like denatured globular proteins (Goff et al. 1996; Rangamani et al. 2014; Qu et al. 2019). These types of proteins adopt a spherical structure, resulting from the inward folding of the hydrophobic parts, while the hydrophilic parts are arranged around the outer surface, whereas fibrous proteins are made up of long polypeptide chains (Borasky 1963; He et al. 2020).

Table 1. Protein Content and Molecular Weight of Protein Concentrates Extracted from Soybean Meal, Shrimp Shells, and Skim Milk Powder (Mary et al. 2024)

Preparation of Polyurethane Adhesives

Polyurethane adhesive formulations were prepared with an estimated isocyanate to hydroxyl functions (NCO/OH) ratio of 1.13 to ensure a complete reaction between the polyol and the isocyanate (Meier-Westhues 2019). The incorporation of proteins was completed by substituting the hydroxyl groups of the polyol with the amine groups of the proteins, resulting in the formation of urea bonds (Kumar and Pizzi 2019). Protein concentrates were incorporated into the polyol and dispersed at 1000 rpm for 3 min with a Dispermat LC30 Dissolver (VMA-Getzmann, Reichshof, Germany) with a 45 mm flat turbine. The substitution was made at 5%, 10%, and 15% in terms of chemical function. Polyol substitution was based on the polyol and protein concentrates’ OH and NH contents. The reference, formulated with the same chemicals as the protein-based adhesives, is represented by the formulation containing 0% protein.

Polyurethane adhesives were prepared using either a petrochemical isocyanate or a combination of 75% petrochemical isocyanate and 25% partially bio-based isocyanate. The extent of isocyanate substitution was determined based on the NCO content of the two isocyanates. In this study, two distinct types of adhesive formulations were used: P for the polyurethane formulated with the petrochemical isocyanate and PB for the polyurethane formulated with the combination of two isocyanates.

When the petrochemical isocyanate was used, the adhesives containing soybean meal, shrimp shells, and skim milk powder protein concentrates are labeled as P-B, P-S, and P-M, respectively. In contrast, when a combination of isocyanates was employed, they are denoted as PB-B, PB-S, and PB-M. The compositions of these adhesives, considering different protein contents, are provided in Table 2. For example, for a 10 g mass of adhesive, the mass of protein concentrate in the formulation is between 0.20 g and 0.80 g, depending on the substitution rate and the protein concentrate.

Table 2. Adhesive Compositions Based on the Isocyanates, Polyol, and Protein Concentrates Used

Adhesives Characterization

Thermogravimetric analysis (TGA)

Thermogravimetric analyses were performed to determine the sensitivity of adhesives to temperature. The temperatures of the maximum degradation rates in the adhesive were determined using a TGA 851e analyzer (Mettler Toledo, Greifense, Switzerland). Approximately 5 mg of adhesive was placed in a 70 µL reusable sapphire crucible and heated from 35 to 800 °C at a 10 °C/min heating rate under nitrogen flow. The analyses were conducted in triplicate.

Gel time

The gel time is the maximum duration during which the system remains adequately fluid to be used as a substrate (Desai et al. 2003). Gel time measurements were conducted using a Gel Time Meter 22A (Sunshine, Philadelphia, PA, USA). Each sample was poured into a test tube and placed in an oil bath at 100 °C. The samples were heated until gel formed, and the time to reach this state was recorded in seconds. These measurements were carried out in triplicate.

Time-dependent viscosity

Viscosity serves as a crucial physical property that influences adhesive behavior. Suitable viscosity provides the adhesive with optimal flow characteristics, simplifies handling to attain strong bonding in the final product, and enables penetration into the initial wood cells, which is crucial for establishing mechanical anchoring (Luo et al. 2016). Viscosity measurements were conducted at 25 °C using a Bohlin Visco88 viscometer (Malvern Instruments Limited, Worcestershire, United Kingdom), taking a viscosity measurement every 5 min. This viscometer adopts a concentric cylindrical design comprising a rotating inner cylinder and a stationary outer cylinder. The inner cylinder has a diameter of 25 mm, while the outer cylinder has a diameter of 27.55 mm. The analyses were performed at a rotational speed of 6.39 rads^-1. The torque developed on the inner cylinder by the sample is directly correlated with the sample’s viscosity and should be in a range of 0.5 to 9.5 mN.m to ensure accurate measurement. Approximately 15 mL of the sample was introduced into the viscometer cylinder. The analyses were performed in triplicate.

Effective penetration

Analyzing adhesive penetration is vital for evaluating the adhesive’s capability to infiltrate the wood surface and pores, thereby ensuring the formation of a robust and durable bond. Two-layer laminated wood panels were prepared for each adhesive and sectioned into 1 cm³ cubes. Cross-sections measuring 40 µm in thickness were precision-cut using a HistoCore AUTOCUT automatic rotary microtome from Leica Biosystems (Buffalo Grove, USA). These sections were mounted on glass slides, immersed in deionized water, covered with glass coverslips, and sealed with nail varnish to prevent water evaporation. The penetration of adhesives into the wood was visually assessed using a VHX-7000 digital microscope at 700x magnification, manufactured by Keyence Co. Ltd. (Osaka, Japan). To facilitate dark field observation, full ring lighting was employed. As described in prior studies, quantitative analysis of adhesive penetration into the wood substrate was performed by measuring the effective penetration depth (EP) (Sernek et al. 1999; Bastani et al. 2016; Qin et al. 2016). The EP quantifies the total adhesive area within the interphase region, divided by the bond line width. It specifically excludes the cell walls and the unfilled lumen area. EP can be calculated using the following Eq. 1,

(1)

where EP is the effective penetration depth (µm), the area of adhesive object i (µm²), and the width of the maximum rectangle defining measurement area (µm). Three measurement zones were studied for each adhesive.

The measurement parameters outlined in Eq. 1 are visually depicted in Fig. 1. Subsequently, ImageJ software was employed to quantify and measure these parameters. The results were analyzed using a least significant difference (LSD) test.

Fig. 1. Measurement parameters in an experimental image

Delamination

Delamination is a common failure mode in adhesive joints. It occurs when the glue bond is not adequate to resist moisture cycling, at the interface between the adhesive and the substrate, in this case, wood, or within the adhesive layer itself. The delamination tests were performed to evaluate cross-laminated wood panels’ bond strength and delamination resistance under specified conditions. For this test, three-layer cross-laminated timber panels were produced using spruce wood (Picea abies Karst). The dimensions of one layer were 11 x 11 x 2 cm³. A mass of 2.34 g of adhesive was deposited per layer. The panels were then pressed under a force of 10 kN at 20 °C for 2 h. A total of six panels were made per adhesive. The panels were conditioned at 22 °C with a relative humidity of 50% for seven days before being cut and subjected to analyses according to Annex A of the NBN EN 16351 (2021) standard. Finally, the percentage of delamination for each specimen was calculated using the following Eq. 2,

(2)

where l_tot,delam is the total length of delamination, expressed in mm, and l_{tot,bonding joint}is the sum of the lengths of all bonding joints in a delamination, expressed in mm. Delamination for each adhesive composition is expressed as the average of the delamination results per specimen. The results were analyzed using an LSD test.

Shear strength

Mechanical analyses were carried out to determine the shear strength of the adhesives used to bond the wood.

Two-layer glued laminated timber panels were produced. The dimensions of one layer were 11 x 11 x 2 cm³. A mass of 2.34 g of adhesive was deposited per layer. The panels were pressed under 10 kN of force at 20 °C for 2 h. A total of two panels were made per adhesive using spruce wood (Picea abies, Karst). Wood specimens were conditioned at room temperature with a relative humidity of 50% for seven days. Subsequently, the panels were cut into specimens with dimensions of 5 x 5 x 4 cm³ for conducting shear strength tests according to Annex D of the NBN EN 14080 (2013) standard.

Before carrying out the shear tests, the specimens were divided into two groups: the first group was subjected to treatment A1 of standard NBN EN 302-1 (2013), i.e., seven days at 20 °C at 65% relative humidity, and the second group to treatment A2 of standard NBN EN 302-1 (2013), i.e., seven days at 20 °C at 65% relative humidity followed by four days of immersion in water at 15 °C. The tests were carried out for the specimens subjected to treatment A2 in the wet state. The load was reported in Newtons (N).

The test results are presented as the arithmetic average of the load of the valid tests, defined as those that exhibit a linear portion of the load-displacement curve. The results were analyzed using an LSD test.

Confocal Raman spectroscopy

Raman analyses were conducted to elucidate the adhesive-wood interaction and enhance comprehension of shear strength results. The analyses were conducted on the same cross-sections used for optical microscopy, using a Senterra II Raman microscope (Bruker Optics Inc., Billerica, USA) equipped with a motorized table (Märzhäuser Wetzlar, Wetzlar, Germany) and an x100 immersion oil objective featuring a numerical aperture of 1.3 (Olympus). The spectra were acquired with a 532 nm laser at 2.5 mW, employing a 250 ms integration time and 8 or 24 co-additions per spectrum. Chemical mapping was performed by scanning an area of 40 x 30 µm with 25 x 25 points.

Table 3. Assignment of Selected Vibrational Bands of Polyol, Proteins, Isocyanate, Urethane, Urea Bonds, and Wood Components

The main characteristic vibrational bands attributed to the various components are displayed in Table 3 (Ukishima et al. 1995; Oliveira Vieira da Cunha et al. 2004; Roohpour et al. 2009; Ghobashy and Abdeen 2016; Ling et al. 2018; Zeng et al. 2023; Cadieux-Lynch et al. 2024).

RESULTS AND DISCUSSION

Thermogravimetric Analysis

Thermogravimetric analyses were performed to determine the thermal resistance of the various adhesives. Figure 2a illustrates the TG curves (thermogravimetric analyses), and Fig. 2b the DTG (derivative thermogravimetric) curves of polyurethane adhesives based on protein concentrates with varying isocyanate compositions. The weight loss of these adhesives can be categorized into two distinct stages. The maximum decomposition temperatures for each stage are presented in Table 4. In the first decomposition phase, observed at approximately 340 °C, the urethane and urea bonds undergo intensified decomposition (Awad et al. 2010; Jiang et al. 2017; Członka et al. 2020; Luo et al. 2020; Jiang et al. 2021). This decomposition phase at 340 °C also corresponds to the decomposition of protein concentrates, which involves the cleavage of the amino acid chains during thermal degradation and is consistent with a previous study (Mary et al. 2024). The second decomposition phase occurs above 450 °C and corresponds to the degradation of the isocyanate hard segments (Jiang et al. 2021).

Notably, the maximum degradation temperatures of the different adhesives remain unaffected by the protein source, isocyanate type, or urea formation, indicating no substantial influence on this parameter.

Fig. 2. TG thermal curves and DTG thermal curves of polyurethane adhesives according to protein sources, protein contents, and isocyanates

Table 4. Maximum Decomposition Temperatures of Polyurethane Adhesives According to Protein Sources, Protein Concentrates Contents, and Isocyanates

Gel Time

The gel time of the different PU adhesive formulations is reported in Fig. 3. Observations show that PB adhesives exhibited longer gel times than P-adhesives. Similar results have been presented in the literature (Malik and Kaur 2018). Indeed, aromatic isocyanates are more reactive than aliphatic isocyanates, due to the negative charge shift towards the direction of the aromatic ring structure (Bengtström et al. 2016). Consequently, the interaction between aliphatic isocyanates and polyol occurs at a slower pace, elucidating the longer gel time observed in PB adhesives compared to their P counterparts (Malik and Kaur 2018).

Fig. 3. Gel time of polyurethane adhesives according to protein sources, protein contents, and isocyanates

Moreover, replacing polyol with protein concentrates leads to a reduction in gel time compared to the references. NH₂, being more nucleophilic than OH due to its easy donation of a lone electron pair and the larger size of the nitrogen atom compared to oxygen, is closer in energy to the transition state, lowering the activation energy and thus accelerating the polymerization reaction (Afagh and Yudin 2010). This reduction is particularly pronounced for PB adhesives.

Given the low molecular weight of M, which ensures a better dispersion into the adhesive matrix, a shorter gel time was expected for P-M and PB-M adhesives compared to those based on other protein concentrates. However, the observed results may be attributed to differences in the protocols for obtaining the various concentrates. While soybean meal and shrimp shells are precipitated at pH 4 after hydrolysis, skim milk is precipitated with acetic acid to pH 3. Literature suggests that a low pH has an impact on nucleophilic groups, which may decelerate urea and urethane formation, resulting in a longer gel time (Maillard et al. 2021). This could explain why the gel times of M-containing adhesives are not shorter than those of the other adhesives in this study (Maillard et al. 2021).

While a prolonged gel time can provide an advantage during the assembly of structural components by allowing sufficient time for application before adhesive curing, it is essential to maintain this duration within reasonable limits to ensure efficient panel manufacturing from an industrial standpoint.

Time-dependent Viscosity

Viscosity measurements were conducted for the various adhesive systems at 25 °C, as illustrated in Fig. 4. These analyses were performed immediately upon mixing the adhesive components.

Notably, the findings indicate a distinct increase in the initial viscosity associated with the inclusion of proteins, regardless of the adhesive type. These results align with established literature, consistently highlighting those introducing fillers, into a polymer system results in heightened viscosity (Schulze et al. 2003; Markovičová, 2021). The greater protein content in M reduces the quantity required to add to the adhesive to achieve the desired protein concentration, thereby reducing viscosity over time. The low molecular weight of its proteins also helps explain these viscosity results. A greater difference was expected between adhesives containing B and adhesives containing S and M due to the presence of β-Conglycinin proteins in B, whose molecular weights are higher than those of S and M (Table 1).

Notable differences emerged between the two adhesive types. The PB adhesives exhibited a longer period to reach maximum viscosity, leading to lower viscosity values for PB adhesives measured at a specific time t compared to P adhesives. It is worth mentioning that BI is designed for ultra-low viscosity applications, which may also explain the difference between PB and P adhesives. In addition, these results are consistent with the gel time findings presented in Fig. 3, which illustrate that PB adhesives exhibit a lengthier gel time, resulting in lower viscosity values for PB adhesives at a given time t compared to P adhesives.

The reduced viscosity of PB adhesives is advantageous, enabling the potential increase in polyol substitution rates with protein concentrates while maintaining viscosity at levels conducive to proper application on wood. This advantage is more challenging to exploit with P adhesives, where higher substitution rates would elevate viscosity to a degree incompatible with effective application in wood bonding processes and industrial practices.

Fig. 4. Viscosity analysis of polyurethane adhesives according to protein sources, protein contents, and isocyanates

Effective Penetration

Detailed analyses were made to explore the effect of integrating proteins into adhesives on the effective penetration depth of these adhesives into wood lumens, as depicted in Fig. 5.

During the penetration process, three primary factors must be considered: the adhesive’s properties, the characteristics of the wood, and the parameters governing the processing (Kamke and Lee 2007). Because the parameters for wood and gluing/pressing process were kept uniform across all adhesives, these factors were not considered when comparing different adhesives. However, the attributes of the adhesive, such as its viscosity and gel time, can influence the extent to which various adhesives penetrate the wood substrate (Kamke and Lee 2007).

It was observed that PB adhesives exhibited greater penetration in the lumens than P adhesives. This aligns with the gel time results (Fig. 3) because PB adhesives, with their longer gel times, have more time available before curing for penetrating compared to P adhesives.

Table 5 shows that PB adhesives with 15% protein content penetrate significantly better into the lumens than the other adhesives in this study. Regarding viscosity, as illustrated in Fig. 4, both B and S result in increased viscosity for both P and PB adhesives. This increase in viscosity contributes to enhanced adhesive penetration into the lumens as protein content increases (Vnučec et al. 2015, 2016). Indeed, low-viscosity adhesives have been demonstrated to flow into the lower part of samples during the pressing process, through the radial bands of parenchyma, reducing the effective penetration into the lumens (Vnučec et al. 2015, 2016). However, the rise in viscosity with increasing protein content is relatively less substantial for P adhesives, resulting in a reduction in lumen penetration compared to PB adhesives.

Fig. 5. Effective penetration of polyurethane adhesives in wood according to protein sources, protein contents, and isocyanates

Table 5. LSD test for Effective Penetration of Polyurethane Adhesives in Wood According to Protein Sources, Protein Contents, and Isocyanates

Different letters indicate significant differences at P ≤ 0.05

Delamination

The effect of isocyanate substitution and the impact of protein content on adhesive performance was investigated for the three different protein sources (Fig. 6).

For P-B and P-S adhesives, increasing protein content helped to reduce delamination, as shown in Fig. 7 and Table 6. The observed results can be explained by the ability of proteins to interact with the isocyanate. Indeed, isocyanates can interact with nucleophilic functional groups, such as hydroxyl and amino groups, present in proteins. The extent of these interactions is directly related to the concentration of proteins in the system, with higher protein content leading to more opportunities for isocyanates to react with the amino groups (Mekonnen et al. 2014). Through these interactions, crosslinked networks are formed, which can improve the adhesive’s bond strength and water resistance of the adhesives (Zeng et al. 2021). In contrast, increasing protein content has no significant impact on P-M adhesives. It has been demonstrated that M protein concentrate has a lower molecular weight than B and S protein concentrates (Table 1). This lower molecular weight leads to less viscous adhesives that penetrate less into the lumens of the wood, which can impact delamination results.

When isocyanates are substituted, adhesives show slightly higher delamination. This may be attributed to the greater crosslinking of P adhesives compared to PB adhesives, due to the greater functionality of PI compared to BI (Kumar and Pizzi 2019).

However, it is noteworthy that the obtained results for P-S and PB-S adhesives align with the findings from the study on adhesive penetration in wood. The superior mechanical anchorage of these adhesives can increase their resistance to delamination. Specifically, increased protein content correlates with enhanced penetration and reduced delamination.

Nevertheless, it is important to note that even the petrochemical reference is not sufficiently resistant to delamination to be used in construction. This is underscored by ASTM D2559 (2018), which stipulates that delamination in any bond line should not exceed 1% for softwoods and 1.6% for hardwoods (ASTM D2559 2018).

Fig. 6. Delamination results of polyurethane adhesives in wood according to protein sources, protein contents, and isocyanates

Table 6. LSD test for Delamination of Polyurethane Adhesives in Wood According to Protein Sources, Protein Contents, and Isocyanates

Different letters indicate significant differences at P ≤ 0.05

Fig. 7. Observed delamination on a wood panel bonded with a) P-S5 adhesive and b) P-S10 adhesive

Shear Strength

Mechanical analyses were performed on P and PB adhesives to assess their mechanical behavior after treatments A1 and A2 (Fig. 8). Table 7 shows that a portion of failures occurred within the adhesive joint in this treatment, and Fig. 9 illustrates the distinction between a 100% failure in the wood and a 100% failure in the adhesive joint.

Fig. 8. Maximum breaking load before the rupture of polyurethane adhesives in wood according to protein sources, protein contents, and isocyanates after A1 and A2 treatments

Table 7. Fracture Percentage in Wood of P Adhesives Based on Protein Content and Isocyanate Type After A1 and A2 Treatments

Table 8. LSD Test for Maximum Breaking Load Before the Rupture of Polyurethane Adhesives in Wood According to Protein Sources, Protein Contents, and Isocyanates After A1 and A2 Treatments

Different letters indicate significant differences at P ≤ 0.05

The statistical analyses, as delineated in Table 8, indicate a noteworthy superiority of P adhesives from treatment A1 over those from treatment A2, except for P-M15. For P adhesives after treatment A1, the addition of proteins decreased mechanical performance compared to the reference adhesive; however, the LSD multiple comparison test proves that this difference was not significant for most P adhesives. When P adhesives were exposed to moisture in treatment A2, the mechanical performance of the adhesives deteriorated. However, the decline in performance was more pronounced for the petrochemical reference adhesive, which exhibited an 80% reduction in maximum breaking load before failure between treatments A1 and A2, meaning that the petrochemical reference is more sensitive to water immersion of the A2 process than protein-based P-adhesives. This can be explained by the fact that urea bonds, present in protein-based polyurethane adhesives, are more stable than urethane bonds (C-O) (Liu et al. 2019).

Fig. 9. Wood samples with 100% wood fracture (top) and 100% adhesive fracture (bottom)

Fig. 10. Optical micrographs of wood samples bonded with a) P-S at 5% protein concentrate content and b) P-S at 5% protein concentrate content. The glue line is marked with GL, the penetrated adhesive is marked with PA, and voids in the glue line are marked with V.

Upon the partial substitution of PI with BI, there was a considerable decrease in the mechanical performance of the adhesives following treatment A1. Although Fig. 5 demonstrates that PB adhesives exhibited greater effective penetration compared to P adhesives, it is essential to emphasize that effective penetration, while important, does not singularly dictate these outcomes. Indeed, PI, having higher functionality than BI contributes to a greater crosslinking density in the adhesive (Kumar and Pizzi 2019). In addition, the presence of aromatic rings contributes to the adhesive’s resistance. Indeed, polyurethanes formulated from aromatic isocyanates have been shown to have better mechanical performances than those formulated from aliphatic isocyanates (Touchet and Cosgriff-Hernandez 2016). This factor may account for the superior performance of P adhesives in comparison to PB. Another comprehensive understanding of the mechanical strength of PB adhesives can also be acquired by investigating the condition of the adhesive glue line. As depicted in (Fig. 10), through microscopic analysis, voids were detected within the glue line of PB adhesives, a characteristic absent in the glue line of P adhesives. These observations can be rationalized by the excessive penetration of PB adhesives into wood lumens, creating voids that weaken the glue line and can affect the adhesive’s mechanical strength. The voids could initiate the rupture mechanism.

Fig. 11. Raman image of wood cell wall of a) P-S at 5% protein concentrate content and b) P-BS at 5% protein concentrate content

Moreover, in the context of PB adhesives, mechanical anchoring, as indicated by effective penetration, holds greater significance, while chemical anchoring is less present, as demonstrated in Raman spectroscopy images (Fig. 11). Raman analyses, performed by following the -C-N bond stretching vibration of urethane and the -N-H bending bond of urea at 1550 cm^-1, confirmed the presence of adhesive within the lumen, as depicted in Figs. 11a and 11b. These specific bands were chosen for the study due to their spectral position, which avoids interference with bands characteristic of wood compounds (Table 3).

Additionally, in the case of P adhesives, the adhesive exhibited reactivity with the wood cell walls, as illustrated by the presence of urethane and urea bonds beyond the lumens in Fig. 11a. The enhanced chemical anchoring of P adhesives to wood cell walls, facilitated by their reactivity, can be attributed to the greater functionality of PI. Additionally, the smaller size of P adhesive molecules promotes chemical reactions with wood cell walls. In contrast, the longer chains of PB adhesives and their lower reactivity, stemming from their isocyanate blend, may contribute to a reduced chemical interaction compared to P adhesives. This distinction might elucidate the observed variance in mechanical performance between these two adhesive types.

Nonetheless, in the case of PB adhesives, even if all failures occurred entirely within the adhesive joint and chemical anchoring was less present than in P adhesives, the breaking load before failure exhibited an increasing trend with an increase in protein content when the adhesives underwent both treatments A1 and A2, which is consistent with the effective penetration of these adhesives shown in Fig. 5. However, these results remained lower than those of P adhesives due to the discontinuity of the adhesive line. In addition, a reduction in the breaking load before failure was also noted between treatments A1 and A2, which is consistent with the observations in P adhesives.

CONCLUSIONS

This study aimed to explore sustainable alternatives to petroleum-based adhesives, with a particular focus on polyurethane adhesives incorporating proteins sourced from industrial byproducts. The investigation delved into the influence of proteins and isocyanates on adhesive performance. To minimize environmental impact, local raw materials derived from non-recyclable or excess industrial waste, such as soybean meal, shrimp shells, and skim milk powder, were selected.
The thermal properties of the adhesives remained unaffected by the protein content and isocyanate system variations, ensuring stability and reliability in these formulations.
The addition of proteins increased the viscosity and the effective penetration of the adhesives but it either decreased or maintained the gel time compared to the petrochemical reference. An equilibrium must be found in formulating P and PB adhesives to harness the optimal properties of each adhesive.
The mechanical characterization of the adhesives underwater exposure revealed a positive correlation between increased protein content and the retention of mechanical strength in adhesives compared to the reference adhesives without protein. However, partial substitution of isocyanate resulted in mechanical strength losses of up to 98%.
While the incorporation of proteins reduces the delamination, particularly for B and S protein concentrates, further studies are still needed to reduce the delamination performance of different adhesives.
Further work needs to be done to optimize the pressing conditions to reach the optimal performances of the structural bond lines.
The partially biobased isocyanate used in this study is an aliphatic isocyanate, and this type of isocyanate produces adhesives with inferior performances. Therefore, it is necessary to explore the development of biobased aromatic isocyanates.

ACKNOWLEDGMENTS

The authors are truly grateful to Vincent Claude, Yves Bédard, Diane Gagnon, Luc Germain, Solène Péllerin, Marie Soula, Sacha Tremblay, and Jérémy Winninger. Special thanks to Antonin Maheux for his help and expertise. The authors would also like to thank Richard Nadeau, Audrey Cote, Serge Lacasse, the members of Buildwise (Limelette, Belgium) and of the Renewable Materials Research Center (CRMR), as well as the Natural Sciences and Engineering Research Council of Canada (NSERC) industrial chair on eco-responsible wood construction (CIRCERB) and its industrial partners.

Funding

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada for the financial support through its IRC, and CRD programs (IRCPJ 461745- 18 and RDCPJ 524504-18) as well as the industrial partners of the NSERC industrial chair on eco-responsible wood construction (CIRCERB). The authors are also grateful to the Ministère de l’Économie et de l’Innovation du Québec through the PSO-I-009 project as well as the Fonds de recherche du Québec – Nature et technologies (project #332793).

Author Contributions

Conceptualization, A.M., P.B., and V.L.; methodology, A.M., P.B., S.P., S.C., and V.L.; software, A.M.; validation, A.M.; formal analyses, A.M.; investigation, A.M.; writing—original draft preparation, A.M.; writing—review and editing, A.M., P.B., S.P., A.H., S.C., and V.L.; supervision, P.B., and V.L.; project administration, P.B., and V.L.; funding acquisition, A.M. and P.B. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest

The authors declare that there are no conflicts of interest. The funders had no role in the design of the study; the collection, analysis, or interpretation of data; the writing of the manuscript; or the decision to publish the results.

REFERENCES CITED

Afagh, N. A., and Yudin, A. K. (2010). “Chemoselectivity and the curious reactivity preferences of functional groups,” Angewandte Chemie – International Edition 49(2), 262-310. DOI: 10.1002/anie.200901317

Al-Jabari, M. (2022). “Waterproofing coatings and membranes,” in: Integral Waterproofing of Concrete Structures: Advanced Protection Technologies of Concrete by Pore Blocking and Lining, Woodhead Publishing, Sawston, UK, pp. 393-435. DOI: 10.1016/B978-0-12-824354-1.00012-X

Arias, A., Entrena-Barbero, E., Feijoo, G., and Moreira, M. T. (2022). “Sustainable non-isocyanate polyurethanes bio-adhesives for engineered wood panels are revealed as promising candidates to move from formaldehyde-based alternatives,” Journal of Environmental Chemical Engineering 10(1), 2213-3437. DOI: 10.1016/j.jece.2021.107053

ASTM D2559 (2018). “Standard specification for adhesives for bonded structural wood products for use under exterior exposure conditions,” ASTM International, West Conshohocken, PA, USA. DOI: 10.1520/D2559-12AR18

Awad, W., Wilkie, C. A., and Awad, W. H. (2010). “Investigation of the thermal degradation of polyurea: The effect of ammonium polyphosphate and expandable graphite,” Polymers 51(11), 2277-2285. DOI: 10.1016/j.polymer.2010.03.033

Bai, Y., Liu, X., Shi, S. Q., and Li, J. (2020). “A tough and mildew-proof soybean-based adhesive inspired by mussel and algae,” Polymers 12(4), article 756. DOI: 10.3390/POLYM12040756

Bastani, A., Adamopoulos, S., Koddenberg, T., and Militz, H. (2016). “Study of adhesive bondlines in modified wood with fluorescence microscopy and X-ray micro-computed tomography,” International Journal of Adhesion and Adhesives 68, 351-358. DOI: 10.1016/j.ijadhadh.2016.04.006

Bengtström, L., Salden, M., and Stec, A. A. (2016). “The role of isocyanates in fire toxicity,” Fire Science Reviews 5(1), article 4. DOI: 10.1186/s40038-016-0013-2

Borasky, R. (1963). “Introduction: Some important aspects of protein fiber research,” in: Ultrastructure of Protein Fibers 1963, pp. 1-3. DOI: 10.1016/B978-1-4832-2838-9.50007-6

Cadieux-Lynch, R., Leroux, E., Hermann, A., Pellerin, S., Keralta, A., Blouin, M., Larouche, J., Profili, J., Chamberland, J., and Landry, V. (2024). “From waste to building material: How whey ultrafiltration permeate can increase wood stability. Journal of Materials Research and Technology 29, 1911-1922. DOI: 10.1016/j.jmrt.2024.01.228

Calle, M., Lligadas, G., Ronda, J. C., Galià, M., and Cádiz, V. (2016). “Non-isocyanate route to biobased polyurethanes and polyureas via AB-type self-polycondensation,” European Polymer Journal 84, 837-848. DOI: 10.1016/j.eurpolymj.2016.04.022

Członka, S., Strakowska, A., and Kairyte, A. (2020). “Application of walnut shells-derived biopolyol in the synthesis of rigid polyurethane foams,” Materials 13(12), article 2687. DOI: 10.3390/ma13122687

Desai, S. D., Patel, J. V., and Sinha, V. K. (2003). “Polyurethane adhesive system from biomaterial-based polyol for bonding wood,” International Journal of Adhesion and Adhesives 23(5), 393-399. DOI: 10.1016/S0143-7496(03)00070-8

Ghobashy, M. M., and Abdeen, Z. I. (2016). “Radiation crosslinking of polyurethanes: Characterization by FTIR, TGA, SEM, XRD, and Raman spectroscopy,” Journal of Polymers 2016, article ID 9802514. DOI: 10.1155/2016/9802514

Goff, H. D., Hill, A., and Ferrer, M. A. (1996). “Milk proteins-caseins; casein micelles; whey proteins; enzymes,” in: Dairy Science and Technology, Ebook, University of Guelph, Ontario, Canada.

Gui, C., Zhu, J., Zhang, Z., and Liu, X. (2016). “Research progress on formaldehyde‐free wood adhesive derived from soy flour,” Adhesives – Applications and Properties 187-200. DOI: 10.5772/65502

He, H., Yang, C., Wang, F., Wei, Z., Shen, J., Chen, D., Fan, C., Zhang, H., and Liu, K. (2020). “Mechanically strong globular‐protein‐based fibers obtained using a microfluidic spinning technique,” Angewandte Chemie 132(11), 4374-4378. DOI: 10.1002/ange.201915262

Husnaeni, Maruddin, F., Malaka, R., and Prahesti, K. I. (2019). “Study on the use of various concentration of acetic acid and different precipitation duration on casein characteristics,” IOP Conference Series: Earth and Environmental Science 343(1), article ID 012035. DOI: 10.1088/1755-1315/343/1/012035

Javni, I., Zhang, W., and Petrović, Z. S. (2003). “Effect of different isocyanates on the properties of soy-based polyurethanes,” Journal of Applied Polymer Science 88(13), 2912-2916. DOI: 10.1002/app.11966

Jiang, S., Shi, R., Cheng, H., Zhang, C., and Zhao, F. (2017). “Synthesis of polyurea from 1,6-hexanediamine with CO2 through a two-step polymerization,” Green Energy and Environment 2(4), 370-376. DOI: 10.1016/j.gee.2017.05.001

Jiang, W., Hosseinpourpia, R., Biziks, V., Ahmed, S. A., Militz, H., and Adamopoulos, S. (2021). “Preparation of polyurethane adhesives from crude and purified liquefied wood sawdust,” Polymers 13(19), article 3267. DOI: 10.3390/polym13193267

Kamke, F. A., and Lee, J. N. (2007). “Adhesive penetration in wood – A review,” Wood and Fiber Science 39(2), 205-220.

Kumar, R. N., and Pizzi, A. (2019). Adhesives for Wood and Lignocellulosic Materials, Scrivener Publishing LLC, Beverly, MA, USA.

Ling, L., Li, J., Zhang, G., Sun, R., and Wong, C. P. (2018). “Self-healing and shape memory linear polyurethane based on disulfide linkages with excellent mechanical property,” Macromolecular Research 26(4), 365-373. DOI: 10.1007/s13233-018-6037-9

Liu, W. X., Yang, Z., Qiao, Z., Zhang, L., Zhao, N., Luo, S., and Xu, J. (2019). “Dynamic multiphase semi-crystalline polymers based on thermally reversible pyrazole-urea bonds,” Nature Communications 10(1), article 4753. DOI: 10.1038/s41467-019-12766-6

Luo, J., Luo, J., Li, X., Gao, Q., and Li, J. (2016). “Effects of polyisocyanate on properties and pot life of epoxy resin cross-linked soybean meal-based bioadhesive,” Journal of Applied Polymer Science 133(17), article 43362. DOI: 10.1002/app.43362

Luo, S., Gao, L., and Guo, W. (2020). “Effect of incorporation of lignin as bio-polyol on the performance of rigid lightweight wood–polyurethane composite foams,” Journal of Wood Science 66(1), article 23. DOI: 10.1186/s10086-020-01872-5

Maillard, D., Osso, E., Faye, A., Li, H., Ton-That, M. T., and Stoeffler, K. (2021). “Influence of lignin’s pH on polyurethane flexible foam formation and how to control it,” Journal of Applied Polymer Science 138(18), article 50319. DOI: 10.1002/app.50319

Malik, M., and Kaur, R. (2018). “Influence of aliphatic and aromatic isocyanates on the properties of poly(ether ester) polyol based PU adhesive system,” Polymer Engineering and Science 58(1), 112-117. DOI: 10.1002/pen.24537

Markovičová, L. (2021). “The effect of filler content on the viscosity of polymer composites,” Conference Quality Production Improvement–CQPI 3(1), 293-298. DOI: 10.2478/cqpi-2021-0028

Mary, A., Blanchet, P., Pepin, S., Chamberland, J., and Landry, V. (2024). “Upcycling of protein concentrates from industrial byproducts into polyurethane wood adhesives,” BioResources 19(1), 1165-1189. DOI: 10.15376/biores.19.1.1165-1189

Meier-Westhues, U. (2019). Polyurethanes: Coatings, Adhesives and Sealants. European Coatings, Vincentz Network, Hannover, Germany.

Mekonnen, T. H., Mussone, P. G., Choi, P., and Bressler, D. C. (2014). “Adhesives from waste protein biomass for oriented strand board composites: Development and performance,” Macromolecular Materials and Engineering 299(8), 1003-1012. DOI: 10.1002/mame.201300402

Oliveira Vieira da Cunha, F., Henrique Roesler Melo, D., Bassanesi Veronese, V., and Madalena Camargo Forte, M. (2004). “Study of castor oil polyurethane-poly(methyl methacrylate) semi interpenetrating polymer network (SIPN) reaction parameters using a 2³ factorial experimental design,” Materials Research 7(4), 539-543.

Pizzi, A., and Mittal, K. L. (2005). Handbook of Adhesive Technology (Third edition), CRC press, Boca Raton, FL, USA.

Qin, L., Lin, L., and Fu, F. (2016). “Microstructural and micromechanical characterization of modified urea-formaldehyde resin penetration into wood,” BioResources 11(1), 182-194. DOI: 10.15376/biores.11.1.182-194

Qu, W., Häkkinen, R., Allen, J., D’Agostino, C., and Abbott, A. P. (2019). “Globular and fibrous proteins modified with deep eutectic solvents: Materials for drug delivery,” Molecules 24(19), article 3583. DOI: 10.3390/molecules24193583

Rangamani, P., Xiong, G. Y., and Iyengar, R. (2014). “Multiscale modeling of cell shape from the actin cytoskeleton,” Progress in Molecular Biology and Translational Science 123, 143-167. DOI: 10.1016/B978-0-12-397897-4.00002-4

Raydan, N. D. V., Leroyer, L., Charrier, B., and Robles, E. (2021). “Recent advances on the development of protein-based adhesives for wood composite materials—A review,” Molecules 26(24), article 7617. DOI: 10.3390/molecules26247617

Roohpour, N., Wasikiewicz, J. M., Moshaverinia, A., Paul, D., Rehman, I. U., and Vadgama, P. (2009). “Isopropyl myristate-modified polyether-urethane coatings as protective barriers for implantable medical devices,” Materials 2(3), 719-733. DOI: 10.3390/ma2030719

Schulze, K. A., Zaman, A. A., and Söderholm, K. J. M. (2003). “Effect of filler fraction on strength, viscosity and porosity of experimental compomer materials,” Journal of Dentistry 31(6), 373-382. DOI: 10.1016/S0300-5712(03)00091-5

Sernek, M., Resnik, J., and Kamke, F. A. (1999). “Penetration of liquid urea-formaldehyde adhesive into beech wood,” Wood and Fiber Science 1, 41-48.

Touchet, T. J., and Cosgriff-Hernandez, E. M. (2016). “Hierarchal structure–property relationships of segmented polyurethanes,” Advances in Polyurethane Biomaterials 3-22. DOI: 10.1016/B978-0-08-100614-6.00001-9

Ukishima, S., Iida, K., Takahashi, Y., and Fukada, E. (1995). “FT-IR observation on orientation of dipolar urea bonds in polyurea prepared by vapor deposition polymerization,” Ferroelectrics 171(1), 351-361. DOI: 10.1080/00150199508018446

Vnučec, D., Goršek, A., Kutnar, A., and Mikuljan, M. (2015). “Thermal modification of soy proteins in the vacuum chamber and wood adhesion,” Wood Science and Technology 49(2), 225-239. DOI: 10.1007/s00226-014-0685-5

Vnučec, D., Kutnar, A., and Goršek, A. (2017). “Soy-based adhesives for wood-bonding–a review. In Journal of Adhesion Science and Technology 31(8), 910-931. DOI: 10.1080/01694243.2016.1237278

Vnučec, D., Mikuljan, M., Kutnar, A., Šernek, M., and Goršek, A. (2016). “Influence of process parameters on the bonding performance of wood adhesive based on thermally modified soy proteins,” European Journal of Wood and Wood Products 74(4), 553-561. DOI: 10.1007/s00107-016-1018-1

Yang, I., Kuo, M., Myers, D. J., and Pu, A. (2006). “Comparison of protein-based adhesive resins for wood composites,” Journal of Wood Science 52(6), 503-508. DOI: 10.1007/s10086-006-0804-5

Zeng, G., Zhu, F., Aladejana, J. T., Zhou, Y., Li, K., Luo, J., Li, X., Dong, Y., Wang, K., and Li, J. (2023). “Barley – A yet un-tapped feedstock for improved vegetable protein-based wood adhesives,” Journal of Materials Chemistry A 11(21), 11310-11325. DOI: 10.1039/d3ta00619k

Zeng, Y., Xu, P., Yang, W., Chu, H., Wang, W., Dong, W., Chen, M., Bai, H., and Ma, P. (2021). “Soy protein-based adhesive with superior bonding strength and water resistance by designing densely crosslinking networks,” European Polymer Journal 142, article ID 110128. DOI: 10.1016/j.eurpolymj.2020.110128

Article submitted: March 18, 2024; Peer review completed: April 6, 2024; Revised version received and accepted: April 8, 2024; Published: April 17, 2024.

DOI: 10.15376/biores.19.2.3520-3542