Lard was rendered in the laboratory and stored at -18 °C for use. Commercial soybean oil was purchased from Wilmar International Limited (Shanghai, China). Potatoes were purchased from a local market. Other chemicals were obtained from Sinopharm Chemical Reagent Co. Ltd. The solvents employed were of AR grade.

All of the reactions were monitored by thin-layer chromatography (TLC) performed on silica gel GF₂₅₄ (homemade). Ultraviolet (UV) spectra were recorded with a UV-2450 spectroscopic instrument (Shimadzu Corp, Kyoto, Japan). The quartz cell was used as a vessel. The protection factors for antioxidants were measured by the rancimat method (Metrohm, Shanghai, China). All NMR experiments were conducted on a Bruker Avance 500 MHz spectrometer (Bern, Switzerland) operating at 500.1 MHz for ¹H and 125.8 MHz for ¹³C, respectively. Chemical shifts were reported in ppm using TMS as internal standard. The single-crystal structure analysis was performed using X-ray diffraction on a Bruker Smart Apex II diffractometer (Germany) at 296K, μ (Cu/Kα)=0.108 mm^-1, Dcalc =1.354 mg/m³, 5584 reflections were measured, 1935 were unique (Rint=0.0147). The structure was dissolved with SheLXS a structure solution program using Direct methods and refined with SheLXL refinement package using least squares minimization (Zhou et al., 2016).

The samples were examined in duplicate. Pearson’s correlation test and t test were conducted using Origins 8 (Origin Lab, MA, United States). Reported data show standard deviation below 5%. Statistical significance has p values always below 0.05.

The compounds 2-4 were prepared according to earlier methods (Saiz-Poseu et al., 2012; Wang et al., 2006) with slight modifications (Scheme 1).

Preparation of 2-(tert-butyl)-6-methoxy-4-methylphenol (2): 7.80 mL of 2-methoxy-4-methylphenol (1) (61.7 mmol, 1 equiv) were added to 85% phosphoric acid (19 ml) and tert-butanol (30 ml, 328.4 mmol, 5.32 equiv) at 90 °C. After stirring for 7 hours, and concentrated to remove most of the tert-butanol under reduced pressure, ethyl acetate (100 mL) and water (10 mL) were added. The organic phase was dried over Na₂SO₄ and concentrated under reduced pressure. The crude products were purified with flash chromatography (petroleum ether:ethyl acetate= 20:0.5) to yield 2, a light, colorless oil (8.50 g, 71%). ¹H NMR (500MHz, CDCl₃, ppm) δ1.49 (s, 9H), 2.37 (s, 3H), 3.93 (s, 3H), 5.91(s, 1H), 6.67 (s, 1H), 6.78(d, J=0.35Hz. 2H). ¹³C NMR (125.8 MHz, CDCl₃, ppm) δ21.53, 29.56, 34.67, 56.14, 109.423, 127.92, 135.24, 142.01, 146.55.

Preparation of 3-(tert-butyl)-4-hydroxy-5-methoxybenzaldehyde (3): 3.00 g of compound 2 (15.5 mmol, 1 equiv) were added to tert-butanol (35 mL), stirred at 0 °C, then liquid bromine (2.5 ml, 48.7 mmol, 3.14 equiv) was added drop-wise, under nitrogen atmosphere for 20 minutes, while keeping the solution pink. After the addition, the mixture was stirred at room temperature for 1h, quenched by 50 ml H₂O and then partitioned with 50 ml ethyl acetate. The organic phase was dried over Na₂SO₄ and concentrated under reduced pressure. The crude products were purified with flash chromatography (petroleum ether: ethyl ether=13:1) to yield 3, a light, yellow solid (2.6 g, 80%), ¹H NMR (500MHz, CDCl₃, ppm) δ 1.44 (s, 9H), 3.97 (s, 3H), 6.64(s, 1H), 7.34 (d, J=1.65Hz, 1H), 7.47(d, J=1.65Hz. 2H). 9.84 (s, 1H). ¹³C NMR (125.8 MHz, CDCl₃, ppm) δ29.14, 34.70, 56.33, 106.79, 125.34, 128.30, 135.65, 147.25, 150.41, 191.52.

Preparation of 3-(tert-butyl)-4,5-dihydroxybenzaldehyde (4): 2.00 g of compound 3 (9.6 mmol, 1 equiv) were dissolved in CHCl₃ (20 mL), and stirred at 0 °C. After, AlCl₃ (1.79 g, 13.5 mmol, 1.4 equiv) was added in batches and then pyridine (3.4 mL, 42.3 mmol, 4.4 equiv) was added drop-wise. The mixture was refluxed for 24 h, cooled down to 0 °C and washed with 10% HCl (60 mL x 3) until the aqueous phase became light blue. It was then was partitioned with 150 ml ethyl acetate. The organic phase was dried over Na₂SO₄ and concentrated under reduced pressure. The crude products were purified with flash chromatography (petroleum ether: ethyl ether=4:1) to yield 4, a white powder (1.4 g, 75%), ¹H NMR (500 MHz, acetone-d₆, ppm) δ 1.46 (s, 9H), 7.31(d, J=1.8Hz, 1H), 7.44 (d, J=1.8Hz, 1H), 9.78 (s, H). ¹³C NMR (125.8 MHz, acetone-d₆, ppm) δ28.71, 34.44, 111.56, 122.79, 128.48, 135.96, 144.99, 150.48, 190.80.

Preparation of BCA: 1.00 g of 3-(tert-butyl)-4,5-dihydroxybenzaldehyde (4) (5.2 mmol, 1 equiv) and malonate (0.54 g) were dissolved in a mixture of benzene (5 ml), pyridine (0.6 ml, 5.2 mmol, 1 equiv) and piperidine (0.06 mL, 0.7 mmol, 0.13 equiv). The solution was stirred at refluxing temperature for 3.5 h, cooled to 25 °C, poured into the mixture of conc. HCl_(aq) and ice water, and then was partitioned with ethyl acetate (20 ml). The organic phase was dried over Na₂SO₄ and concentrated under reduced pressure. The crude products were purified with flash chromatography (petroleum ether: ethyl ether=3:2) to yield 5 (BCA) (1.04 g, 85%), which was re-crystallized from methanol to afford needle-like light beige crystals, ¹H NMR (500 MHz, acetone-d₆, ppm) δ 1.45 (s, 9H), 6.24 (d, J=15.85Hz,1H), 7.10(d, J=1.85Hz, 2H), 7.57 (J=15.85Hz, 1H). ¹³C NMR (125.8 MHz, acetone-d₆, ppm) δ28.86, 34.41, 111.39, 114.19, 120.16, 125.08, 136.00, 144.79, 147.13, 147.00, 168.07.

Commercial soybean oil was purified by silica gel column chromatography to remove its endogenous antioxidants. Different 0.02% (w/w) antioxidants were added to soybean oil samples (500 g) separately and then the oil samples were heated to 180 °C. Fresh potato slices, about 2±0.5 mm thickness, were added at the rate of 20 g/h, fried for 8 minutes, then removed from the oils. Oil samples were tested every 3 h during the frying period until 60 h. Acid values (AV) and Iodine values (IV) were determined for each sample using the International Union of Pure and Applied Chemistry (IUPAC) method (Paquot and Hautfenne, 1987).

The CD values were calculated from the absorbance and the final concentration of the samples (g/100 mL). The results were expressed as conjugated diene (CD) values and computed as follows (Glende and Recknagel, 1994; Zuta et al., 2007):

Where, A is absorbance of the sample at 233 nm.

C is final dilution concentration of the sample (g/100 mL).

P is the length of the measuring UV cell (cm).

The antioxidant capacities of BCA, TBHQ and caffeic acid were measured by the rancimat method (Metrohm, Shanghai, China) at different concentrations in oils. An air flow (20 L/h) was bubbled through the oil heated at different temperatures: 80, 100 and 120 °C. The tests were carried out in duplicate. The protection factors (Pf) were calculated according to the following formula (Silva et al., 2001):

Where, IP_x = the induction period of the sample in the presence of antioxidant.

IP₀ = the induction period of the sample without antioxidant.

A 0.5 mL methanol solution of different antioxidant concentrations was added to a DPPH-methanol solution (3.0 mL, 40 mg/L), and allowed to react for 30 min at 25 °C, then, absorbance at 517 nm was recorded at different time intervals on a UV visible spectrophotometer. The DPPH concentration in the reaction medium was calculated from the calibrated curve. The radical scavenging activity (%) was calculated according to the following formula (Liu et al., 2014; Piang-Siong et al., 2017):

A DPPH solution (3 mL) with methanol (0.5 mL) was used as negative control. A solution (3 mL) with methanol (0.5 mL) with a different concentration of antioxidant was used as a blank. EC₅₀ (i.e., efficient concentration of the substance that produces 50% scavenging), obtained for the different antioxidants was calculated by linear regression of plots, where radical scavenging activity (%) was plotted against concentration.

The ¹H and ¹³C chemical shifts of compound BCA agree well with its chemical structure, which is compared with the spectra data for caffeic acid reported in the literature (Jeong et al., 2011).

The single-crystal structure analysis of BCA was performed using X-ray diffraction. The crystal data of BCA presented (Figure 1; Table 1) give the perspective views of this compound together with its atomic labelling system. Thus, it can be confirmed for the absolute configuration of BCA ((E)-3-(3-(tert-butyl)-4,5-dihydroxyphenyl)acrylic acid).

During the oxidation of lipids containing methylene substituted dienes and polyenes, a shift in the position of the double bond was observed, due to isomerization and conjugated bond formation (conjugated dienes). The percentage changes of CD during frying (Figure 2, A), were as follows: BCA>Caffeic acid>TBHQ>Blank.

No obvious change was observed during the initial 12 hours (Figure 2, B). But, later, the change in AV manifested the capacity of antioxidants as follows: CBA>Caffeic acid>TBHQ>Blank.

When soybean oils were in deep frying at 180 °C, some polyunsaturated fatty acids became conjugated, and some double bonds were destroyed, thus, decreasing iodine values. IV changes at 180 °C (Figure 2, C) increased as follows: BCA>Caffeic acid>TBHQ>Blank.

According to their Pf values (Figure 3), TBHQ, Caffeic acid and BCA showed good antioxidant capacities in the concentration 0.02% at 80, 100, and 120 °C, as follows: BCA>Caffeic acid>TBHQ>Blank.

In different concentrations, 0.01, 0.02, and 0.04% at 100 °C (Figure 4), Pf values changed as follows: BCA>Caffeic acid>TBHQ>Blank. BCA, however, showed a remarkable Pf value (i.e., 79.9) at a concentration of 0.04% at 100 °C.

The scavenging DPPH radical is widely used to evaluate antioxidant capacity due to its simple, rapid, sensitive and reproducible procedure. In this study, kinetic investigation was performed to estimate the speed of DPPH–Antioxidant reaction. On the other hand, EC₅₀ of TBHQ, caffeic acid and BCA were measured in terms of hydrogen-donating or radical-scavenging capacity.

The EC₅₀ values for the reducing power of TBHQ, Caffeic acid and BCA were 0.1690, 0.1283, and 0.1748 μg/L, respectively. The reducing power decreased as follows:

TBHQ≈BCA>Caffeic acid. The ease of donating hydrogen (Figures 5, A and B) was Caffeic acid>TBHQ≈BCA.

In the deep frying oil and rancimat tests, BCA showed stronger antioxidant capacity than TBHQ and caffeic acid (Scheme 2):

In the DPPH system, BCA showed approximate antioxidant capacity with TBHQ, and lower capacity compared to caffeic acid:

EC₅₀ values. The steric effect of the tert-butyl group hinders free radical DPPH from occupying carbon positions on the benzene ring of BCA. As a result, the free radical scavenging activity of BCA is less than caffeic acid and TBHQ in the DPPH system. However, the synergistic effect of the catechol group benefited the leaving out of hydrogen atoms, and the large conjugated system of the acrylic acid group enabled the stabilized radical resonance to capture more DPPH radicals, thus, making the free radical scavenging activity of BCA similar to TBHQ in the DPPH solution.

Kinetic behavior. The steric effect of the tert-butyl group on the BCA caused the second phenol hydrogen in the ortho-tert-butyl group to remain after the first one left. The same response was observed for the oxidation of compound 3 to compound 4 without protecting the phenol hydrogen.