#### 3.1. Resin Properties

The synthesis of MESS (

Figure 2) was catalyzed using AMC-2, which also suppresses the hydroxyl-epoxy side reactions. The inhibitor, hydroquinone, was used in order to prevent premature polymerization. The final resin was fully characterized via acid number titration,

^{1}H NMR, FTIR, GPC, and rheometry. The results are summarized in

Table 1.

The structure of the resin was characterized via FTIR and it is shown in

Figure 4. The band around 3480 cm

^{−1} was characteristic absorption of the O–H stretching while 1747 cm

^{−1} was that of the C=O stretching of the fatty acid chains. The bands around 1718, 1637, and 941 cm

^{−1} were assigned to the absorptions of C=O stretching, C=C stretching, and =C–H out of plane bending of methacrylates, respectively. The characteristic band for the oxirane C–O is absent after methacrylation. The summarized assignments are presented in

Table 2.

The structure of MESS was further characterized by

^{1}H NMR and results are presented in

Figure 5. The protons from the starting epoxy are typically between 2.8–3.2 ppm. As seen in

Figure 5, those peaks are very minimal. Meanwhile, new peaks around 6.1(H

_{a}), 5.6(H

_{b}), and 1.9(H

_{c}) ppm appeared. These correspond to the protons in the vinyl and methyl groups of methacrylate. These results confirm the structure of MESS and indicate successful conversion of the oxiranes to methacrylates. Detailed analysis of the resin can be found in the published work of Yan and Webster [

15].

#### 3.2. Mechanical Properties of the Composite

Composite specimens were prepared, as explained in

Section 2.2; the resulting fiber volume fraction of the manufactured composite was 35%.

Results of frequency sweeps are presented in

Figure 6. It is worth mentioning that, because

tan $\mathsf{\delta}$ is the ratio of loss modulus to storage modulus, it has not been plotted in the graphs. A horizontal shifting of the storage modulus values was performed using TA Instruments Data Analysis software. To perform the shifting process, the storage modulus curve at 30 °C was picked as the reference curve, and all other curves were shifted to the left. In this process, the loss modulus curves were also shifted with the same values of shift factors. The resulting plots are presented in

Figure 7.

According to TTS assumptions, the same shift factors should be valid for all viscoelastic parameters [

21,

34]. Based on

Figure 7, although a smooth master curve is obtained for the storage modulus, the curve for loss modulus is not satisfactory. This is indication of the fact that only one set of horizontal shift factors is not enough for all three sets of curves.

TA Instruments Data Analysis software was employed to move curves simultaneously and shift them both horizontally and vertically. A two-dimensional minimization method was used by the software and storage modulus, loss modulus and tan

$\delta $ curves were simultaneously moved horizontally and vertically until they superpose. In addition to auto generated shift factors, an excel spreadsheet was used to modify the horizontal and vertical shift factors manually. The resulting master curves are presented in

Figure 8. Much smoother master curves are obtained with this approach. Based on these results it is valid to conclude that MESS resin reinforced with flax fiber is a thermorheologically complex material and, to generate a smooth master curve, both horizontal and vertical shift factors are necessary [

35].

Horizontal shift factors used in

Figure 7 are plotted in

Figure 9a. Horizontal and vertical shift factors used in

Figure 8 are plotted in

Figure 9b. The dependency on temperatures of shift factors, both for horizontals and vertical shift factors, complies with the Arrhenius equation, which has the following form [

1]:

where

${a}_{T}$ is the shift factor,

T_{r} (K) is the reference temperature and

T (K) is an arbitrary temperature at which horizontal shift factor

${a}_{T}$ is desired. In this equation,

Q is the activation energy (kJ/mol) and

R is the universal gas constant (J/mol·K).

Based on

Figure 9, the corresponding values for

Q could be calculated using Equation (6). The calculated value for activation energy is 47.52 (kJ/mol) considering only horizontal shift factors of

Figure 9a. If both horizontal and vertical shift factors of

Figure 9b are considered, the values of

Q are calculated to be 55.48 (kJ/mol) and 42.95 (kJ/mol) based on horizontal shift factors and vertical shift factors, respectively.

Figure 10 shows creep data collected at different temperatures. Similar to the work of other researchers for natural fiber/thermoplastics [

36] and natural fiber/thermosets [

1], who have only applied horizontal shift factors, and neglected the application of shift factors to other viscoelastic properties, strain curves are horizontally shifted in reference to the strain curve at 30 °C to generate the creep master curves presented in

Figure 11. As observed, an acceptably smooth master curve is obtained by this method.

Horizontal shift factors obtained from shifting the storage modulus curve in

Figure 7 were applied. The results are presented in

Figure 12. As can be seen, the creep curves do not superimpose and no satisfactory master curve is generated. Once more, this is an indication that horizontal shift factors are not solely sufficient to generate a master curve. In the next step, both horizontal and vertical shift factors obtained from shifting storage and loss modulus in

Figure 8 is used to shift creep data. The result is a master curve shown in

Figure 13. A smooth master curve is obtained by this method. In addition, by comparing the master curve obtained by horizontal shifting of creep data, with the curve obtained with horizontal and vertical shifts, it is perceived that the latter covers a wider range on the time axis.

As mentioned before, a long-term creep test was performed at 30 °C for 24 h to check the validity of the obtained master curves. Creep data at 30 °C were used to find the parameters in Findley Power Law and Nutting Power Law. The parameters then were used to extrapolate the creep data to 24 h. Extrapolated curves based on Findley Power Law and actual creep data for 24 h are presented in

Figure 14. At longer times, there is deviation between Findley model and actual creep data. However this model over estimates the strain creep values therefore provides more conservative values of creep strain. On the other hand, Nutting Power Law has a much better estimate of the creep data and stays closer to actual data.

Figure 15 shows the comparison of the actual creep data with two master curves generated with horizontal shift factors, and horizontal and vertical shift factors. In both curves, there is a deviation from the actual creep data at longer times and both master curves tend to underestimate the creep strain.

Behera

et al. [

37] reinforced soy milk based composites with jute to investigate the properties of these composites and the viability of their use in packaging, furniture and automotive industries. They studied non-woven and woven jute soy composites using series of mechanical testing and characterized, strength, flexibility, hydrophilicity, surface characteristics, and susceptibility to biological degradation. Results of their study showed that developed composites are a feasible option for housing and office space applications as well as automotive industry. In another study, O’Donnell

et al. [

38] used vacuum assisted resin transfer molding (VARTM) to reinforce a soy-oil based resin (acrylated epoxidized soybean oil) with flax fiber with different fiber volume fractions, and studied the mechanical and thermal properties of the composites. They found that 33.3% styrene content in the resin was the optimum ratio of styrene to provide the maximum composite properties while maintaining the minimum styrene content possible. They run Dynamic Mechanical Analysis on their samples and measured loss and storage moduli of their composites. Similar to Behera

et al. [

37], their findings revealed that manufactured composites are suitable for applications in housing and automotive industries.

There have been other studies on developing vegetable oil-based or soy-based resins to be used in composites [

39,

40,

41,

42,

43]. Most of these studies have focused on the synthesis or curing mechanism of bio-based resins; however, to the best of the authors’ knowledge, there have not been any studies on creep behavior or application of TTS methods to flax fiber reinforced soy-based thermosets. The authors believe that the findings and results of current study will be a great contribution in this area and valuable for further developments of their application in more engineering and structural applications.