Comparative Analysis on the Epoxidation of Soybean Oil using Formic and Acetic Acids

Abstract

The objective of this study was to make a comparative analysis of the effect of formic and acetic acids as oxygen carriers on the epoxidation of soybean oil used with hydrogen peroxide as the oxygen donor. Comparative analysis between the use of formic acid (FA) and acetic acid (AA) was studied to obtain the most effective oxygen carrier that yielded high oxirane oxygen contents (OOC). The epoxidation reaction was carried out using a stoichiometric ratio of 1:0.5:1, 1:0.5:0.5, and 1:0.5:2 of soybean oil: formic/acetic acid: hydrogen peroxide. The synthesis was performed at three reaction times (2, 4 and 6 h) at a constant temperature of 50°C. Samples prepared using FA and AA were characterized using ASTM D1652-11 and confirmed by Fourier transform infrared (FTIR) spectroscopy. The result of this study proved FA to be an effective oxygen carrier compared to that of AA based on the high OOC and percent yield achieved. The optimum epoxidized soybean oil (ESO) sample using FA was obtained at a reaction time of 6 h using 2 moles of H₂O₂, yielding an OOC of 7.45 at a relative conversion to oxirane of 98%. Samples of FA were further characterized to prove the optimum parameters that gave the highest OOC using rheology and gel permeation chromatography (GPC). Rheology data revealed an increase in the viscosity that implied an increase in the degree of epoxidation. GPC indicated an increase in the molecular weight at low reaction times, then a decrease resulting in a change in the structure of the triglyceride and consequently an increase in the extent of epoxidation.

1. Introduction

A number of industries focus their attention on epoxy resins as an acceptable matrix (resin) system. Epoxy resins are known for their excellent mechanical and thermal properties ¹ , good chemical resistance and the ability to bond well with reinforcing fibers ² in fiber-reinforced composites. However, most resins offering excellent mechanical properties are derived from petroleum by-products; hence, they are neither ecofriendly nor bio-degradable. This provides an impetus to synthesize a bio-based epoxy resin system from vegetable oils.

The use of vegetable oils is one of the most inexpensive ways to create a resin system. Vegetable oils are sustainable, renewable, biodegradable with low toxicity, and for many of the vegetable oils, they are the most abundant biological feedstock available in large quantities ³ . The methods to obtain polymers from vegetable oils include direct polymerization through the double bonds or other reactive functional groups, chemical modification of the double bonds that introduces reactive functional groups, or chemical transformation to reduce fatty acid chains to monoglycerides ⁴ . Among these, chemical modification of the double bonds by way of the epoxidation mechanism is the most widely used method. This epoxidation method has demonstrated itself to be the most conventional and effective process used to produce polymers from vegetable oils, primarily because it is inexpensive and well-controlled^5,6. The epoxidation method converts the double bonds in the vegetable oil to epoxide groups, thus going from a less reactive to a more reactive system because of the highly strained oxirane ring (epoxide). The presence of an epoxide group in vegetable oils makes for a good monomer for the replacement of the petrochemical diglycidyl ether of bisphenol A (DGEBA).

Epoxidation can be carried out in a number of ways that include the conventional chemical treatment; use of acid ion exchange resins (AIER), enzymes, or a metal catalyst ³ . Among all methods listed, the conventional chemical treatment method and the method using acid ion exchange resins are the most preferred ones. Both methods use carboxylic acid (formic acid or acetic acid) that acts as the oxygen carrier and hydrogen peroxide as the oxygen donor. It is important to know that hydrogen peroxide is considered to be the most promising oxidizing agent because it is cheap and pollution-free ⁷ . However in regard to the carboxylic acids, due to acetic acid being a weaker acid compared to formic acid, acetic acid is normally used with a catalyst which results in the acid ion-exchange resin method. This acid ion-exchange method uses an insoluble gel type catalyst such as Amberlite IR-120H ³ .

Reaction of the carboxylic acids (formic acid or acetic acid) and hydrogen peroxide together forms peracids (performic acid (PFA) or peracetic acid (PAA), respectively), that constitutes the aqueous phase of the system. The peracid further reacts with the double bonds in the oil to produce epoxide groups ⁸ in the oil phase. Thus, it is important to achieve a high oxirane oxygen content (OOC), which simply indicates the number of epoxide groups that will result in increased crosslinking and consequently increased properties. The ability to achieve desirable OOC requires precise optimization of the system resulting in reduction of production time and cost. However, there are a number of factors that can affect the epoxidation reaction, such as reaction time, reaction temperature, amounts of formic or acetic acid, amount of hydrogen peroxide and stirring speed^9–12.

Monono et al. ¹³ carried out a synthesis reaction of epoxidized sucrose soyate while optimizing the reaction temperature (55, 60, and 65°C), moles of acetic acid (0.25, 0.375 and 0.5), moles of H₂O₂ (1, 1.5 and 2), catalyst amounts (1.5, 3.75 and 6 g), and reaction times (3.5, 4.5 and 5.5 h). A conversion rate of 95.9% from double bonds to epoxide groups was obtained when using 0.375 mol of acetic acid, 2 mol of hydrogen peroxide, and 6 g of catalyst for 4.5 h at 60°C through the American Oil Chemists Society (AOCS) method Cd 9-57. Similarly, other researchers have performed various tests to optimize the above mentioned parameters^{10,12,14–16}. Optimized parameters varied with each study because researchers used different oils and hence had variation in fatty acid percentage. The major difference in optimizing parameters in the epoxidation of vegetable oils stems from the ability to determine the most effective oxygen carrier. Formic acid, being highly reactive, was chosen by Gupta et al. ¹⁷ and He et al. ¹⁸ , while others found acetic acid to be useful along with Amberlite IR-120H^3,8,19.

Hence, this research focuses on determining the effect of varying the reaction time and molar ratios of hydrogen peroxide using both formic acid and acetic acid to synthesize an epoxidized soybean oil (ESO) system. The overall goal is to understand how the OOC changes because of these factors using two different peracids.

2. Experimental

2.1 Materials

Soybean oil used in this study was obtained from Sigma Aldrich in its raw form. The fatty acids and the subsequent percent composition of each fatty acid of soybean oil are shown in Table 1 . This information was provided by Sigma Aldrich. Formic acid, acetic acid and hydrogen peroxide were also obtained from Sigma Aldrich. Amberlite IR-120H was obtained from Acros Organics™.

Table 1

Fatty acid percentage composition of soybean oil (Sigma Aldrich, USA)

Fatty Acids	Percent Composition (%)
Lauric	0.2
Myristic	0.1
Palmitic	9.8
Stearic	2.4
Arachidic	0.9
Palmitoleic	0.4
Oleic	28.9
Linoleic	50.7
Linolenic	6.5
“unknown”	0.1

Other materials used in this study consist of sodium bicarbonate, sodium sulfate and distilled water used for neutralization and drying of the epoxidised oil to obtain the final product. Additionally, tetraethylammonium bromide (anhydrous), perchloric acid (60 %), acetic anhydride, methylene chloride, crystal violet indicator and potassium acid phthalate were used in the titration method to determine the OOC based on ASTM D1652-11.

2.2 Synthesis of Epoxidized Soybean Oil (ESBO)

Epoxidation of soybean oil was carried out by placing 1 mol of soybean oil and 0.5 mol of formic acid in a three-neck flask equipped with a thermometer and magnetic stirrer. The flask containing the solution was submerged in an oil bath on top of a magnetic stirring hot plate. The mixture was stirred at 900 rpm at 50°C. Once the desired temperature was achieved, the temperature was reduced by 5°C before the required amount of hydrogen peroxide was added. This reduction in the temperature was to avoid a drastic temperature rise as a result of the exothermic nature of reaction during the addition of hydrogen peroxide. With the temperature reduced by 5°C, the required amount of hydrogen peroxide was slowly added to the mixture using a titration tube to start the epoxidation reaction. After the completion of the addition of hydrogen peroxide, the reaction was allowed to continue until the end of reaction time. This exact procedure was performed for the varied reaction times (2, 4 and 6 h) and different ratios of soybean oil to hydrogen peroxide (1:1, 1:1.5 and 1:2 mol ratio).

The reaction using acetic acid was carried out in the same manner with the addition of Amberlite IR-120H added during the addition of oil and carboxylic acid. Amberlite IR-120H was used in the amount of 20 wt. % of oil. A schematic representing the process used to perform the synthesis of epoxidized soybean oil is shown in Figure 1 .

Figure 1

Schematic representing the synthesis of epoxidized soybean oil

The epoxidized oil was allowed to cool to room temperature after the completion of the reaction. The post treatment of the synthesized oil involved filtering and washing the oil several times with distilled water to remove any excess acid and separate the oil phase from the water phase. The reaction using acetic acid and Amberlite IR-120H was first filtered through a filter paper to separate the bead-like catalyst from the epoxidized oil before washing with distilled water. Sodium bicarbonate in the amount of 5 wt. % dissolved in 100 mL of distilled water was placed in a titration tube and used to neutralize the reaction to obtain a pH of 7. The epoxidized oil solution was placed in a beaker on a magnetic stirring plate to allow for constant mixing, while the solution of sodium bicarbonate was titrated into the beaker. The pH was checked regularly using pH strips. The neutralized epoxidized soybean oil was then dried over sodium sulfate in the amount of approximately 20 wt. % of the undried epoxidized oil and placed in an oven overnight at 70°C.

2.3 Characterization Techniques

Oxirane oxygen content (OOC) and epoxy equivalent weight (EEW) were calculated according to ASTM D1652-11. Fourier transform infrared (FTIR) spectroscopy was used to determine the chemical bonds and change in functional groups. This test was carried out using Shimadzu FTIR 8400s setup equipped with MIRacle TM ATR. Each sample was scanned from 400 to 4000 cm⁻¹ with a resolution of 4 cm⁻¹. The key goal for FTIR was to verify the presence of epoxide groups as a result of varied molar concentrations of hydrogen peroxide and varied reaction times.

For FA samples, rheological studies were carried out using TA Instruments AR-2000 equipped with environmental testing chamber (ETC) and 25 mm parallel plates. Scans were performed at room temperature with varying shear rate from 0 to 300 s⁻¹. Each test recorded the change in viscosity as a function of the shear rate in order to determine processing and performance conditions. The results for rheology were further analyzed using TA Instruments Data Analysis software. Gel permeation chromatography studies were carried out at North Dakota State University using Waters Corporation: Module system with Refractive Index detector. Tetrahydrofuran (THF) was used as the mobile phase at a flow rate of 1 mL/min, and columns were maintained at 40°C. The samples were first filtered then added to a 1.5 mL glass vial with THF at a concentration of 1 mg/mL. Astra V software was used to calculate the molecular weight.

3. Results

3.1 Determination of OOC using Titration method based on ASTM D1652-11

The oxirane oxygen content (OOC) is an important parameter in the characterization of epoxidized vegetable oils and is used to determine the number of oxirane groups formed in place of double bonds. Theoretically, the OOC value of the soybean oil used in this study was found to be 7.58. The theoretical OOC (OOC_t) was calculated based on equations used by Cooney ²⁰ . These equations consisted of calculating the theoretical iodine value (IV_t) shown in equation (1) that is used in the theoretical OOC (OOC_t) equation shown in equation (2).

$$I{V}_t={\displaystyle \sum \left[100\times \frac{(253.81)\times (No.\hspace{0.17em}of\hspace{0.17em}Double\hspace{0.17em}Bonds)}{M{M}_{FA}}\right]}$$

(1)

$$OOC=\left[\frac{\left(\frac{IV}{2\times A_i}\right)}{100+\left[\left(\frac{I{V}_t}{2\times A_i}\right)\times A_0\right]}\right]\times A_0\times 100$$

(2)

IV_t was calculated based on the percent fatty acid composition of soybean indicated in Table 1 as well as the molecular weight of each fatty acid in soybean oil. MM_FA is the molar mass of each fatty acid, as 253.81 is the atomic weight of two iodine atoms that are ideally consumed in one double bond ²¹ . The IV_t was found to be 130.03 (g/100g) respectfully. A_i is defined as the atomic mass of iodine (126.9) and A_o is the atomic mass of oxygen (16) used in equation (2). Thus, the aim of optimizing different parameters was to arrive at OOC values that are closer to this theoretical value result in high relative conversion to oxirane (RCO).

Figure 2 reveals the OOC of the synthesized ESO at various reaction times and molar concentrations of hydrogen peroxide when using FA and AA. The OOC ranged from 1.08–7.45 and 0.67–6.03 respectfully for samples using FA and AA at different reaction times and molar concentrations of hydrogen peroxide. The results using FA reveal a variation in the OOC at 2 and 4 hours with increasing molar concentration of hydrogen peroxide up to 2 mol. An excess of hydrogen peroxide at half ratio (i.e. 1.5 mol) revealed a lower OOC compared to molar concentrations at 1 and 2 mole ratio that affected the rate of epoxidation as shown in other literature ²² . It is suggested that an excess of hydrogen peroxide (especially at a low reaction time of 2 h) does not undergo sufficient diffusion-control or mass transfer of reactants. This may be the result of limited PFA that remains in the aqueous phase and not transferring to the oil phase to initiate epoxide formation ²³ . However, higher OOC was obtained at 4 h compared to 2 h with increasing molar concentration of hydrogen peroxide. Low OOC at 2 and 4 h with increasing molar concentration of hydrogen peroxide is possibly due to degradation of the epoxide groups. This degradation (side reactions) may result from hydrolysis (introducing water molecules), the generation of glycerol ²⁴ , or regeneration of excess FA. Hydrolysis occurs due to the formation of PFA when mixing FA and hydrogen peroxide in the aqueous phase. PFA reacting with the double bonds in the oil leads to the regeneration of excess FA that hinders the course of the reaction. Thus, the mentioned side reactions may indicate an instability of the oxirane rings ²⁵ resulting in the varied oxirane oxygen contents.

Figure 2

Comparison of OOC values of epoxidized soybean oil using formic and acetic acids at different reaction times and H₂O₂ concentration

A more stable reaction occurs by increasing the reaction time to 6 h with increasing molar concentrations of hydrogen peroxide resulting in an increase in the OOC. It is believed that more time is needed to break the double bonds in the oil and generate PFA to create higher epoxide content compared to values at 2 and 4 h. Therefore, the optimum epoxidized soybean oil using FA at 50°C was achieved when using a reaction time of 6 h at a 2 mol ratio of hydrogen peroxide. These optimum parameters yield an OOC of 7.45.

On the other hand, results from the epoxidation reaction of soybean oil using AA and Amberlite IR-120H revealed a variation in the OOC with increasing reaction time and increasing molar concentration of hydrogen peroxide. It is believed that the variation in the OOC values is affected by the concentration and use of Amberlite IR-120H catalyst. Most of the degradation occurs due specifically to the external surface of the Amberlite IR-120H catalyst that results in oxirane ring opening^26,27. The beadlike structure of the catalyst acts as an adsorbent in the presence of AA and other aqueous solutions including water that is generated from the production of PAA; hence, resulting in swelling of the catalyst. Therefore, AA was found to be a less effective and stable oxygen carrier compared to FA.

On the other hand, the epoxide equivalent weight (EEW) using both FA and AA was seen to have an inverse relationship with the OOC. The EEW is defined as the weight of resin in grams, which contains 1 g of the equivalent ²⁸ . This parameter is important in determining the amount of hardener to use when curing the resin. A high epoxide equivalent weight indicates low epoxide groups formed. Consequently, a high epoxide equivalent weight indicates less curing agent needed due to less reactivity of the epoxidized oil which indicates less functional epoxide groups. Therefore, the amount of curing agent needed for the optimum epoxidized soybean oil system (50°C for 6 h using 2 mol of hydrogen peroxide) would be close in weight to the grams of resin being used.

Table 2 presents the RCO and EEW of each epoxidized soybean oil samples. The RCO which indicates the percent of double bonds converted to oxirane groups shows a direct proportionality to the OOC. The first two columns in Table 2 are associated with data obtained for samples using FA, while the last two columns are associated with AA. The optimum parameters obtained when using FA reflects a 98% conversion. This high conversion indicates that there are possibilities to develop a sustainable product of great value.

Table 2

EEW and RCO values of epoxidized soybean oil with varied molar ratios of hydrogen peroxide and reaction time

Sample	Formic Acid		Acetic Acid
	(EEW), (g/eq)	RCO, (%)	EEW, (g/eq)	RCO, (%)
50°C - 2 hours - 1M	431.25 ± 44.59	25.24 ± 2.59	1095.62 ± 155.13	10.80 ± 1.46
50°C - 2 hours - 1.5M	762.18 ± 13.51	14.18 ± 0.25	1228.67 ± 145.36	8.88 ± 1.08
50°C - 2 hours - 2M	556.48 ± 21.11	19.44 ± 0.73	146.81 ± 1.82	73.62 ± 0.91
50°C - 4 hours - 1M	217.77 ± 4.18	49.64 ± 0.95	152.26 ± 1.98	70.99 ± 0.92
50°C - 4 hours - 1.5M	396.21 ± 14.38	27.30 ± 1.00	135.94 ± 4.47	79.56 ± 2.66
50°C - 4 hours - 2M	259.05 ± 2.06	41.72 ± 0.33	170.73 ± 2.11	63.31 ± 0.78
50°C - 6 hours - 1M	126.89 ± 1.53	85.18 ± 1.03	547.12 ± 30.86	19.80 ± 1.15
50°C - 6 hours - 1.5M	116.22 ± 0.54	92.99 ± 0.43	126.52 ± 1.67	85.43 ± 1.12
50°C - 6 hours - 2M	110.00 ± 2.13	98.27 ± 1.90	129.28 ± 2.99	83.63 ± 1.94

3.2 Fourier Transform Infrared (FTIR) Spectroscopy Studies

FTIR was used to verify the formation of epoxide groups. The data obtained showed a number of functional groups associated with the epoxidation reaction as listed in Table 3 . Based on the various functional groups listed, the oxirane (epoxide) functional group is of great importance for epoxidation reactions. The epoxide group was represented by twin peaks formed at 822 and 839 cm⁻¹. Figures 3 (a) and (b) give a deeper insight into the formation of epoxide group for samples prepared using FA and AA. With increasing reaction time and increasing molar concentration of hydrogen peroxide, the intensity of the characteristic peak representing the epoxide group increases. This increase in the intensity reveals the extent to which amount of epoxide bonds exist. Compared to raw soybean oil, as the epoxide group is formed, the peak also gets broader. The broadening of the epoxide peaks suggests that some chemical interaction occurred possibly due to strain and dimer formation. This phenomenon of chemical interaction due to strain is likely to occur as the three-member ring epoxide groups results in weak bonds which leads to high reactivity. However, in Figures 3 (a) and (b), there is a shift to lower frequencies as epoxide groups are formed. This shift may be attributed to formation of free fatty acids ²⁹ , and consequently the structural change of the soybean oil as reaction time and hydrogen peroxide concentration are changed. The results based on the intensity of the peaks around 822 and 839 cm⁻¹ further correlates with data obtained from ASTM D1652-11. Additionally, signs of no epoxide peak gives reasoning for the introduction of side reactions and occurrence of oxirane cleavage that degrades the epoxide rings^12,14,30,31. Further comparison of Figures 3 (a) and (b) reveal three epoxidized samples using FA ( Figure 3(a) ) that reflects in the formation of epoxide group, while AA gave six epoxidized samples ( Figure 3(b) ); suggesting that AA is more complex to the epoxidation reaction than FA. The complexity of AA gives rise to the use of porous Amberlite IR-120H that absorbs an unknown amount of peracetic acid; thus resulting in varied OOC and high EEW values.

Figure 3

FTIR spectrum highlighting the presence of epoxide groups at 822 cm⁻¹ using (a) formic Acid and (b) acetic Acid

Table 3

FTIR functional groups and corresponding frequency peaks present in epoxidized soybean oil samples

Functional Group	FTIR Wavenumber, cm−1
OH stretching	3530
C-H stretching	3012
C-H stretching (CH3)	2920
C-H stretching (CH2)	2850
C=O stretching	1738
C-O-C asymmetric bending	822

Other peaks observed from the FTIR spectrum are 3530 cm⁻¹ (OH stretching), 3012 cm⁻¹ (C-H and C=C stretching) 2920 cm⁻¹ (C-H stretching of CH₃), 2850 cm⁻¹ (C-H stretching of CH₂), and 1738 cm⁻¹ (C=O stretching vibration of carbonyl ester) are listed in Table 3 . The frequency peak at 3530 cm⁻¹ that corresponds to the hydroxyl (OH) group stretching represents side reactions that disturb the oxirane oxygen content (such as hydrolysis, generation of glycerol, and epoxide cleavage). The disadvantage of the OH stretching indicates ring opening of the epoxide ring. FTIR peak at 3012 cm⁻¹ seen in Figures 4 (a) and (b) for epoxidized samples using FA and AA corresponds to the level of unsaturation that disappears as the double bonds are converted and used^32,33. This peak signifies the progress in epoxidation reaction that correlates with the epoxide peak around 822 and 839 cm⁻¹. In addition, peaks at 2920, 2850, and 1738 cm⁻¹ are due to the carbonyl group in the glyceride moiety. Azhar et al. ²⁹ reports the presence of carbonyl groups to be a result of hydroxylation due to chain scission reactions.

Figure 4

FTIR spectrum highlighting the disappearance of C=C at 3012 cm⁻¹ using (a) formic acid and (b) acetic acid

3.3 Rheological and Molecular Weight Studies of Samples using Formic Acid

With FA proven to be the most effective oxygen carrier compared to acetic acid, this section focuses specifically on using this acid to verify the trend obtained from ASTM D1652-11 and FTIR. Rheology tests were performed on the samples at room temperature at shear rates from 0 to 300 s⁻¹. The results of the viscosity versus shear rate of samples synthesized using FA at 50°C with varying reaction time and molar concentration of hydrogen peroxide are shown in Figure 5 . Viscosity study is essential to determine processing conditions of each system for composite manufacturing processes. Initial shear rate reveal a variation in the viscosity indicated by the sharp peaks at 25, 38, and 56 s⁻¹. This sharp peak is indicative of bombardment of the molecules as a result of shearing which results in thixotropic behavior. This thixotropic behavior is associated with the molecules in the oils (whether in the raw form or epoxidized) breaking down and regaining their internal network structure as stress is reduced and removed as a result of initial shear rate ³⁴ . At the end of the thixotropic behavior, the oils come to a defined viscosity where non-Newtonian behavior occurs. The behavior is no longer associated with the molecules, but now reflects either non-Newtonian behavior of shear-thickening or shear-thinning, or Newtonian behavior in the manner of the samples flow. It should be noted that lower viscosities are shown for systems that exhibit low oxirane contents and high molecular weights (or high epoxide equivalent weights). Samples prepared using a reaction time of 6 hours with hydrogen peroxide contents of 1, 1.5 and 2 mol exhibit shear-thickening behavior which relates to structural rearrangement ³⁵ . This shear-thickening is reflected for samples with high oxirane contents that exhibit low epoxide equivalent weights. On the other hand, the other samples resulted in Newtonian behavior indicating independence of the flow rate of the ESO samples with increasing shear rate. At high shear rates, a lower viscosity was observed beyond the viscosity of raw soybean oil for ESO sample synthesized using 1 mole of hydrogen peroxide for 2 h. This low viscosity is an indication of the presence of water molecules as a result of hydrolysis that hindered the degree of epoxidation.

Figure 5

Rheology plot of epoxidation reaction using formic acid at 50°C

Figure 6 shows the GPC plot of the ESO samples using formic acid with varying reaction time and molar concentrations of hydrogen peroxide. Corresponding numerical values of the average molecular weight and polydispersity index (PDI) are listed in Table 4 . The raw soybean oil revealed an average molecular weight of 917 g.mol⁻¹, similar to values reported in the literature^36,37. The epoxidized soybean oil samples tested ranged in molecular weights from 1255 to 1431 g.mol⁻¹. Lower reaction times reveal the highest molecular weights, although there were slight variations with the results compared to those from the results obtained from ASTM D1652-11. These high molecular weight samples indicate that there was no direct change in the structure of the triglyceride³⁸ as confirmed by the disappearance of the epoxide group in FTIR AT 822 cm^–1 and the low oxirane oxygen content obtained from ASTM D1652-11. As the reaction time and molar concentrations of hydrogen peroxide increased, there was a decrease in the molecular weight, which as mentioned earlier suggests higher formation of epoxide groups. This decrease in the molecular weight suggests that triglyceride structure slightly changed with the addition of epoxide groups³⁸. Further analysis of the polydispersity index (PDI) shown in Table 4 confirms the hypothesis mentioned above indicating an increase in PDI as a result of oligomers forming from the generation of epoxide groups. It can be seen that epoxidized soybean oil synthesized at a longer reaction time of 6 h revealed higher PDI. At this point, the reaction temperature plays a key role in the increased PDI, not necessarily due to the presence of epoxide groups, but rather the presence of side reactions from excess water or acid as revealed from ASTM D1652-11 and FTIR. Thus, GPC gives insight and complements results obtained from other characterization techniques for use of FA as an effective oxygen carrier.

Figure 6

GPC plot of major molecular peak using formic acid at 50°C

Table 4

GPC Results showing average molecular weight and PDI as a result of varied molar ratio of hydrogen peroxide and reaction time of epoxidized soybean oil using formic acid

Sample	Average Molecular Weight, (g/mol)	Polydispersity Index (PDI)
Raw Soybean Oil	917	1.073
50°C - 2 hours - 1M	1431	1.007
50°C - 2 hours - 1.5M	1419	1.007
50°C - 2 hours - 2M	1426	1.006
50°C - 4 hours - 1M	1399	1.008
50°C - 4 hours - 1.5M	1417	1.007
50°C - 4 hours - 2M	1423	1.007
50°C - 6 hours - 1M	1295	1.033
50°C - 6 hours - 1.5M	1295	1.035
50°C - 6 hours - 2M	1279	1.036

4. Conclusions

This study focused on the effects of reaction time and molar concentration of hydrogen peroxide on the synthesis of epoxidized soybean oil using FA and AA. The system was optimized by varying the reaction times at 2, 4 and 6 h with varied molar concentrations of hydrogen peroxide at 1, 1.5 and 2 mol to yield high conversions of epoxide groups. Compositional properties to determine the amount of double bonds converted to epoxide groups (oxirane oxygen content (OOC)) was characterized using the titration method based on ASTM D1652-11. Further characterization focused on the verification of the presence of epoxide formation using Fourier-transform infrared (FTIR) spectroscopy. Based on results from ASTM D1652-11 and FTIR, further characterization was performed using rheology and gel permeation chromatography for the carboxylic acid that yielded the most effective results.

It was revealed that FA was proven to be the most effective oxygen carrier compared to AA based on the high percentage yield and OOC obtained. Samples using FA revealed complete decomposition of the reactants compared to the regeneration of Amberlite IR-120H used with AA. As a result, samples using AA resulted in varied OOC with increasing reaction times and molar concentrations of hydrogen peroxide. Low OOC resulted in the presence of side reactions such as hydrolysis, generation of glycerol and oxirane cleavage. Overall, optimum conversion to oxirane was achieved using FA at 6 h using a 2 mol ratio of hydrogen peroxide at 50°C. Based on results from ASTM D1652-11, the optimum system yielded an OOC of 7.45 and a RCO of 98%. This result was verified using Fourier-transform infrared (FTIR) spectroscopy shown at frequencies of 822 cm⁻¹. Rheology and GPC verified the results from ASTM D1652-11 and FTIR for samples synthesized using FA. High viscosities were revealed for samples that resulted in high degree of epoxidation. Similarly, a high average molecular weight and low PDI at short reaction times of 2 and 4 h reveal little to no change in the structure of the triglyceride. Lower average molecular weight and high PDI signifies the presence of oligomers as a result of the formation of epoxide groups. With FA being the most effective oxygen carrier, this work is in its initial stages to characterize and synthesize a bio-based epoxy resin system for use in natural fiber-reinforced composites.