*2.1. Manufacturing of the Samples*

The PET layer was applied on to the WG material (supplied by Lantmännen Reppe AB of Stockholm, Sweden with a gluten protein content of 77.7 wt %) using two different application methods. The first method involved compression molding (Fortijne Presses TP 400, Barendrecht, Netherlands) WG powder at 150 ◦C for 30 min at 290 kN force. Next, 15 wt % of Jeffamine (Jeffamine EDR 148 procured from Huntsman Corporation, Göteborg, Sweden) was brushed onto the molded WG surface and then compression molded again with the initially 250 μm thick PET films (Mylar® A, Synflex Insulation Systems, Blomberg, Germany) using the same pressing conditions. The amount of Jeffamine added was decided based on previous studies [19,20]. This sample was named "Brushed" since the Jeffamine was applied externally. In the other application method, 15 wt % Jeffamine was first added to the WG powder, then mixed through dry blending. This was followed by flash freezing in liquid N2 and grinding (Retsch GmBH, 5657, Haan, Germany). The WG powder containing the Jeffamine was then compression molded along with the PET sheets using the same pressing conditions as mentioned previously. This sample was named "Ground". The neat WG sample (as the control) was also manufactured using the same process as for the "Ground" sample, but without the Jeffamine. The mold dimension was 100 × <sup>100</sup> × 4 mm3, thereby yielding samples of the same sizes. The PET layer was attempted to be forcibly and manually removed to qualitatively gauge its adhesion. It was observed that the PET was completely stuck to the WG polymer. For the WVTR measurements, the mold was circular with a diameter of 70 mm and a thickness of 1 mm. The PET layers were applied to both sides of the samples.

#### *2.2. Testing of the Samples:*

The tensile strengths and moduli (chord moduli between 0.05% and 0.25% strain) of the laminated and control samples were determined on an Instron Universal Testing Machine (10 kN load cell) (Model No. 5566, Instron, Norwood, MA, USA), using specimens that were 4 mm (the WG part) thick, 12.7 mm wide, and 100 mm long. The samples were milled out from the produced plates. The gauge length was 50 mm and the crosshead speed used was 50 mm/min, following the ASTM D638 Protocol [21].

Three top-down samples and three cross-sectional samples were prepared for nanoindentation in this study. The top-down samples were cut into squares with a ca. 15 mm side length and the cross-sectional samples were cut into squares with a ca. 10 mm side length using a SiC cut-off wheel (Struers, Cleveland, OH, USA). The cross-section of the cross-sectional samples was ground with 240 and 600 grit-size SiC abrasive papers (Struers) and polished with a 6 μm diamond suspension (Struers). All samples were then cleaned with ethanol and fixed onto a mild steel plate using a cyanoacrylate adhesive. Figure 1 shows the orientation of the samples for the nanoindentation tests.

**Figure 1.** Sample orientation for nanoindentation.

The nanoindentation procedure was performed using a Hysitron TI950 TriboIndenter (Bruker, Billerica, MA, USA) at the University of Auckland, New Zealand. For both procedures, a 1 μm diamond conical tip, commonly used for the nanoindentation of non-metallic materials, was used. On the top-down samples, five sets of nine standard quasi-static indents (5 s loading, 2 s holding, and 5 s unloading) with increasing maximum loads from 1000 to 9000 μN were done at five random locations on the surface of the PET layer. On the cross-sectional samples, two sets of nine indents with a fixed maximum load of 5000 μN were done at the interfacial region of the PET layer and the WG substrate. The hardness and modulus were calculated from the load-displacement data and the details are available elsewhere [22].

The WVTR of the samples were determined using 25 cm2 VF2201 permeability cups from TQC. The cups, with a volume of 20 mL water, were fitted with the circular samples (exposing 25 cm2 of the film to water vapor) and then placed in a room with 50% relative humidity at 23 ◦C. The mass loss due to water vapor transmission was measured according to ASTM D1653 protocol [23]. The water vapor transmission rate was calculated based on the water loss data between the third and fourth day.

A goniometer (CAM 200, KSV Instruments Ltd., Espoo, Finland) was used wherein a droplet of distilled water was placed on the flat surface of the sample and its contact angles were measured.

Fourier transform infrared spectroscopy (FT-IR) of the samples was performed in ATR mode on a Perkin Elmer Spectrum 100 instrument, Waltham, MA, USA. The layered samples were placed over the crystal to comprehend the chemical changes in that region. For each sample, 64 scans were obtained and averaged from the 600 to 4000 cm−<sup>1</sup> wavelength with a resolution of 4 cm−1. The scanning electron microscopy (SEM) of the tensile fractured samples and of the interfacial (PET/WG) regions were conducted on a Hitachi TM 100 table top SEM, Tokyo, Japan (10 kV voltage, 6 mm working distance). All the tests mentioned in this manuscript were performed in replicates (tensile tests had five replicates, WVTR had two replicates, and the goniometer had ten replicates). The standard deviation measured was reported in the error bars of the figures and as numeric values in the table.

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

#### *3.1. Mechanical Properties*

#### 3.1.1. Tensile Properties

The tensile properties of the laminated samples and the neat sample are summarized in Table 1. It can be observed that the neat WG had the highest tensile strength amongst the tested samples. Its tensile strength value of ca. 29 MPa is similar to that of the tensile strength of neat polypropylene [24], which is an important economic consideration. It is an essential first step in developing bio-based plastics that have the same tensile strength as a common and popular synthetic polymer (here polypropylene). The processing condition for the compression molding was decided after trial tests and a pressing temperature of 150 ◦C for 30 min was the most desirable in terms of good adhesion between PET and WG, and good cohesion of the WG matrix (enabled through sufficient melting during the molding cycle). It should be noted that a polyethylene film was also tested for adhesion to WG, but the effect observed with PET was absent. It was also observed that without pressure during the molding operation and without the diamine, the adhesion between PET and WG was absent. In addition, the particular processing condition enabled a heat-induced cross-linking (disulfide cross-linking) of the gluten network, which is known to enhance the mechanical strength of WG-based materials [25,26].

The addition of the diamine to WG caused some changes in the protein, which consequently reduced the tensile strength, ductility/strain at break and toughness. The effects were largest for the sample where the diamine had been mixed with the WG (Ground) and less for the sample with the diamine applied (Brushed) at the WG surface. It is known that at high temperatures, nitrogen-containing compounds such as diamine depolymerize polymers that possess amide groups [27]. The modulus, however, did not decrease in the presence of the diamine. An estimation of the laminate modulus employing the rule of mixtures for parallel geometry (using the neat WG modulus, Table 1 and PET modulus, 2.7 GPa) gave an unrealistically low value compared to those in Table 1. This revealed that the modulus of the WG component was higher in the presence of diamine due to the cross-linking [22,28]. During the tensile testing, no delamination of the PET layer was observed in any of the samples, which reaffirmed the fact that the bonding of PET to the WG polymer was strong. The results showed that in terms of mechanical properties, the brushing method was superior to the diamine/WG mixing (Ground) method. It should be noted though that both the Brushed and the Ground materials had sufficient mechanical properties for many applications (e.g., packaging, encapsulation of pest- or weed-control agents, toys, covers, frames for electrical devices, household appliances, and furniture) [13,29].


**Table 1.** Tensile properties.

#### 3.1.2. Nanoindentation

The use of the nanoindentation technique has been previously proven to be effective in evaluating the interfacial (amongst other) properties of polymer composites [22,30,31]. Therefore, in this study, the same technique was applied to determine the nanomechanical properties of the PET laminated and neat WG samples (sample thickness of 4.2 mm). Figures 2 and 3 show the nanomechanical hardness

and reduced modulus of the tested top-down samples, respectively. The hardness was essentially independent of the depth of the indentation and on the same level for both PET-layered WG samples, which was expected as the indentation was always in the PET layer. The indentation depth ranged from ca. 300 nm (for a 1000 μN load) to 2000 nm (for a 9000 μN load). This result suggests that the measured hardness was that of the PET with little to no substrate effect(s). For neat WG, there seemed to be a small but consistent increase in hardness inwards into the sample, however, based on standard deviation, the values were not significantly different.

**Figure 2.** Nanoindentation hardness measured top-down.

From Figure 3, it can also be observed that the reduced modulus for neat WG increased with increasing depth, while the reduced modulus for the PET-layered samples remained depth-independent. Evidently, on a submicron to micron-range, the WG surface was less stiff than the interior, but still at a level comparable to that of the PET layer. From the nanoindentation hardness and reduced modulus values, it can be stated, that although the use of the diamine could have some adverse effects on the tensile strength and extensibility (see Section 3.1.1), the PET layers could increase the wear resistance of the samples, which is desirable (Figure 4). The wear resistance can be calculated as the ratio of the nanoindentation hardness to the modulus [32]. Figure 4 shows that the laminated samples had a significantly higher wear resistance when compared to that of the neat WG, and that the wear resistance of the laminated samples were the same.

The load vs. displacement curves of the samples tested top-down are presented in Figure 5. The unloading curves did not overlap the loading curves due to energy loss (damping) and a certain degree of plastic deformation. These effects were similar for the two PET-layered plates, but larger for the neat WG. Figure 6 shows the nanoindentation scan area of the top-down tested samples where the impression left after the conical-tip had indented the samples is clearly visible. This shows that the samples had deformed not only elastically, but also plastically, as also observed in Figure 5. Additionally, we noticed that the PET surfaces were flatter/smoother when compared to the WG material, implying a more elastic surface.

**Figure 3.** Reduced nanoindentation modulus measured top-down.

**Figure 4.** Wear resistance measured top-down.

**Figure 5.** Load vs. displacement curves measured top-down.

**Figure 6.** Nanoindentation regions (before and after indentation) measured top-down. The width of 50 μm in the images are the lateral dimension (*x*-*y* axes). The color bar to the right of each image represents the height of the sample surfaces.

To comprehend the mechanical properties of the interface between the PET and WG, the samples were also analyzed by nanoindentation in the cross-section. The nanomechanical properties of the interfacial region are critical to gauge the nature of the adhesion of the laminations. The nanoindentation regions of the cross-sectional specimens are displayed in Figure 7 (optical image). Figure 8 shows that the hardness of the area containing the interface of the Ground sample (0.18 GPa) was similar to that of the neat WG (0.19 GPa) and the PET (0.21 GPa). The area containing the interface of the Brushed sample had a significantly lower hardness (0.07 GPa) than the corresponding area of the Ground sample. The reduced modulus (Figure 9) was also significantly lower in the area containing the interface of the Brushed sample. Hence, the interface of the Brushed sample was both weaker and less stiff than in the Ground sample. This suggests that the adhesion between the PET layer and WG was stronger in the Ground sample when compared to the Brushed sample. This can be attributed to the method of PET application that created a higher cross-link density. In the Brushed sample, the PET was in direct contact with the diamine, which had an aminolytic effect, resulting in the loss of PET structural integrity (see Sections 3.4 and 3.5).

**Figure 7.** Nanoindentation regions of the cross-sectional samples.

**Figure 9.** Nanoindentation modulus (reduced) of the cross-sectional samples.

*3.2. Resistance Towards Water and Its Vapor*

#### 3.2.1. WVTR Tests

The water mass loss curves during the permeation measurement are given in Figure 10. The initial slow loss of water in the case of the WG film was due to an extensive uptake of water vapor in the film. Figure 11 shows how the initially flat film expanded and deformed during the test. The WVTR of the WG film (1 mm thick) was 79.1 ± 10.7 g·m−2·day−<sup>1</sup> (calculated between the third and fourth day). The WVTR of the neat PET film was, as expected, significantly lower: 11.2 ± 0.6 g·m−2·day−1. The Ground sample had a WVTR of 15.6 ± 4.2 g·m−2·day−<sup>1</sup> that was similar to the neat PET film value (within the standard deviation). The measured WVTR of the brushed sample was, however, higher (23.3 ± 4.9 g·m−2·day<sup>−</sup>1). The water vapor transmissions normalized to (WVTR values × total PET thickness) the actual PET thickness were 2800 ± 150, 3060 ± 820, and 4570 ± 960 g·μm/(m2·day) for the neat PET, Ground, and Brushed samples, respectively. It was noted that there was a reduction in the total thickness (on both sides) of the PET (resulting thickness: (98 ± 5) × 2 = 196 μm) in the pressed material when compared to that of the neat PET (250 μm). The somewhat higher transmission rate of the Ground sample could be attributed to the change in the layer thickness. The changes in PET thickness were probably not the cause for the higher WVTR of the Brushed sample. It is more likely that the increase in WVTR was due to the migration of the diamine into the PET under the hot pressing. Not all diamine reacted with the PET (or the WG) and this low molar mass component may serve as an internal plasticizer of the PET component. In addition, the diamine increased the hydrophilicity of the PET layer. The migration was obviously higher for the Brushed samples as all the diamine was already available at the WG–PET interface, whereas it needed to migrate from the matrix to the PET in the Ground samples. Nevertheless, the Ground sample showed an almost 80% reduction in WVTR whereas the Brushed sample showed a reduction of 70% (compared to neat WG).

**Figure 10.** Water content inside the cup as a function of time.

**Figure 11.** State of the neat and PET-layered WG film when exposed to saturated water vapor (100% RH) on the inside of the cup and 50% RH on the outside.

#### 3.2.2. Contact Angle

The low contact angle at the outer PET-layered surface of the Brushed plate (35◦ ± 4◦) verified that the polar and hygroscopic diamine had penetrated the entire PET layer (Figure 12). The PET film had a contact angle of 90◦ ± 1.2◦ and the more polar WG film had a value of 64◦ ± 1◦. The value of the PET-layer on the Ground sample was 68◦ ± 2.5◦. A possible explanation for this lower value compared to the neat (un-pressed) PET film may be because the PET film surface becomes slightly oxidized/hydrolyzed during the pressing operation, however, on a level small enough not to be observed by IR spectroscopy (see below).

**Figure 12.** Water droplets on the samples during the goniometer test.

#### *3.3. Infrared Spectroscopy*

Figure 13 shows that the infrared (IR) spectrum of the Ground sample (measured top-down) was similar to that of the neat PET, showing that no diamine was present in the PET outer surface. The relative size of the PET ester peak (1712 cm−1) to that of the reference peak (ca. 1410 cm−1, phenylene-ring vibration) [33] was also the same (3.3) as that of un-pressed (neat) PET, indicating that the oxidation or hydrolysis of the PET layer was low enough not to be observed in the IR spectrum. On the other hand, the spectrum of the PET layer of the Brushed sample showed a more pronounced absorbance in the 2800–3000 cm−<sup>1</sup> region than that of neat PET due to the presence of the diamine. The 1712/1410 peak ratio was also lower (2.7) than those in the un-pressed PET and Ground samples, indicating a reaction between the diamine and PET.

**Figure 13.** FT-IR spectra of the different neat components and the PET laminated WG.

In order to assess the strength of the bonding between the PET and WG, an attempt was made to peel off the PET layer from the WG and to run IR on the side of the layer that had faced the WG material. Figure 13 reveals that the inner surface (i.e., Peeled PET) showed clear features of WG (observed by the large amide I peak at 1700–1580 cm−1). Additionally the NH2/OH regions in the peeled PET was higher than that of the neat PET. This shows that during the peeling, the fracture went mainly through the WG material and not the PET/WG interface, which indicated a good bond between the PET and WG. Note also that this was the laminated WG with a weakest interface, i.e., the Brushed sample (weaker than that of the Ground sample).
