1. Introduction

Margarines and related water-in-oil emulsions are susceptible to oxidative deterioration during processing and storage. Lipid oxidation generates hydroperoxides and subsequently volatile products that impair flavour, nutritive quality and shelf life. The interfacial organization of emulsions, the distribution of pro- and antioxidants, and interactions with the aqueous phase can make oxidation behaviour differ substantially from that of bulk oils. [1,2,3,4]

Underutilized tropical oilseeds provide complementary lipid resources for formulated fats. Watermelon (Citrullus lanatus), moringa (Moringa oleifera) and African bush mango (Irvingia gabonensis) yield oils with distinct physicochemical and fatty-acid characteristics, while watermelon seeds may also be processed into a milk-like aqueous dispersion. [6,23,24,25,30]

Plant essential oils and botanical extracts are investigated as natural sources of compounds capable of slowing lipid oxidation. Basil (Ocimum basilicum) essential oil is chemically variable, and its activity therefore depends on chemotype, concentration and food matrix. Linalool, estragole, 1,8-cineole and other terpenoids commonly contribute to its profile, but their relative abundance varies with origin and growing conditions. [5,7,8,9,10,11,12,13,14]

The present work aimed (i) to establish the GC-FID and GC–MS profile of basil leaf essential oil obtained in Brazzaville, Republic of the Congo, and (ii) to compare the peroxide- and acid-value trajectories of multioil margarine prototypes made with water or watermelon-seed milk and containing no antioxidant, 0.02% basil essential oil, or 0.02% ascorbic acid.

2. Experimental

2.1. Raw Materials

Watermelon (C. lanatus) seeds originating from Madingou (Bouenza Department) were purchased at Total Market in Bacongo, Brazzaville. Moringa (M. oleifera) seeds were obtained in Brazzaville, and African bush mango (I. gabonensis) kernels originating from Ouesso (Sangha Department) were purchased at Ouenzé Market. Fresh basil (O. basilicum) leaves were collected at Nzaba in the Madibou district of Brazzaville.

2.2. Preparation of Oils, Watermelon-Seed Milk and Essential Oil

The three lipid materials were obtained by mechanical pressing and filtered through muslin. The I. gabonensis kernel paste was gently warmed before pressing; the M. oleifera and C. lanatus seeds were pressed directly. For the aqueous dispersion, 300 g of watermelon seeds were soaked for 16 h, heated at 100–110 °C for 10 min, ground with 1 L of water and filtered through muslin.

Fresh basil leaves were sorted, washed, drained and hydrodistilled. The distillate was extracted with diethyl ether; the organic phase was dried over anhydrous sodium sulfate, filtered and stored in an amber container until analysis.

2.3. GC-FID Analysis

Fresh basil leaves were sorted, washed, drained and hydrodistilled.

Quantitative profiling was performed on an Agilent 6890 gas chromatograph fitted with a DB-5 capillary column (20 m × 0.18 mm i.d.; film thickness, 0.18 μm) and a flame-ionization detector. The oven was held at 50 °C for 3.2 min and then programmed to 300 °C at 10 °C min⁻¹. The injector temperature was 280 °C. Hydrogen was used as the carrier gas at 1 mL min⁻¹; detector gas flows were 40 mL min⁻¹ hydrogen and 450 mL min⁻¹ air. Relative percentages were obtained by peak-area normalization. Two injections were performed.

2.4. GC–MS Analysis and Compound Identification

Qualitative analysis was performed with an Agilent 7890 gas chromatograph coupled to an Agilent 5975 mass spectrometer and equipped with a DB-5 column of the same dimensions. The oven was held at 50 °C for 3.2 min and then programmed to 300 °C at 8 °C min⁻¹; the injector was maintained at 280 °C. Electron-impact ionization was operated at 70 eV and helium carrier gas at 0.9 mL min⁻¹. Components were assigned by comparing mass spectra and reported retention indices with Adams, NIST 2008, MassFinder and the LEXVA Analytique laboratory databases. [18,19,20,21,22]

2.5. Initial Physicochemical Measurements

Peroxide value (PV) and acid value (AV) of the pressed oils were determined by standardized titrimetric procedures and expressed as meq O₂ kg⁻¹ and mg KOH g⁻¹, respectively. The measurements were performed in triplicate. The pH and titratable acidity of the watermelon-seed milk were also measured; titratable acidity was expressed in Dornic degrees (°D). [15,16,17]

2.6. Margarine Formulation and Storage

The fat blend contained 24.14% I. gabonensis oil, 43.83% M. oleifera oil and 13.05% C. lanatus oil (81.02% total). The aqueous phase was water or watermelon-seed milk (17%). Egg-yolk-derived emulsifier was included at 1.63–1.65%, sodium chloride at 0.35%, and antioxidant treatment at 0, 0.02% basil essential oil or 0.02% ascorbic acid. The six formulations are summarized in Table 1.

Table 1. Experimental design for the six margarine prototypes

The fat and aqueous phases were combined at 60 °C under magnetic stirring for 15 min. The mixture was emulsified using an electric beater, cooled on ice to induce crystallization, kneaded to obtain a smooth product, packed in airtight 100-g plastic containers and stored under ambient laboratory conditions for 49 days. PV and AV were recorded weekly.

2.7. Data Treatment

Physicochemical measurements are reported as mean ± standard deviation (n = 3). Essential-oil percentages are the mean of two GC-FID injections. Weekly storage values are presented descriptively because independent batch replication was not documented in the dataset. Final-day percentage changes were calculated against the corresponding antioxidant-free matrix control. No inferential statistics were applied to the storage series.

3. Results and Discussion

3.1. Initial Characteristics of the Lipid and Aqueous Ingredients

The pressed oils differed in their initial oxidation state (Table 2). C. lanatus oil had the lowest PV (3.10 ± 0.29 meq O₂ kg⁻¹), whereas M. oleifera oil had a PV of 12.83 ± 0.65 meq O₂ kg⁻¹. AV varied less, from 2.96 ± 0.42 to 3.73 ± 0.44 mg KOH g⁻¹. These values describe the materials before blending and therefore provide the relevant baseline for interpreting the formulated system.

Table 2. Initial physicochemical characteristics of pressed oils and watermelon-seed milk

3.2. Essential-Oil Composition and Chemotype

The GC-FID profile comprised 42 constituents and represented 99.998% of the detected peak area (Table 3). Estragole (40.665%), linalool (36.079%) and 1,8-cineole (12.757%) dominated and together accounted for 89.501%. The next most abundant constituents were camphor (1.566%), limonene (0.989%), trans-α-bergamotene (0.931%), β-pinene (0.859%) and epi-α-cadinol (0.757%) (Figure 1).

Table 3. GC-FID and GC–MS composition of Ocimum basilicum leaf essential oil

Figure 1. Major constituents (≥0.50%) of O. basilicum leaf essential oil. Values are mean relative peak areas from two GC-FID injections

By chemical class, phenylpropanoids represented 40.665%, oxygenated monoterpenes 52.037%, monoterpene hydrocarbons 3.708%, sesquiterpene hydrocarbons 2.397%, oxygenated sesquiterpenes 1.126% and other compounds 0.065%. The co-dominance of estragole and linalool supports classification as an estragole–linalool type. This profile is consistent with the pronounced chemical polymorphism reported for basil oils, in which genotype and environmental conditions influence the balance between phenylpropanoids and terpenoids. [10,11,12]

3.3. Development of Peroxide Value During Storage

PV increased throughout storage in every formulation (Table 4 and Figure 2A). At day 49, the water-based control reached 35 meq O₂ kg⁻¹, compared with 25 meq O₂ kg⁻¹ for the essential-oil treatment and 15 meq O₂ kg⁻¹ for the ascorbic-acid treatment. Thus, relative to the water-based control, basil essential oil and ascorbic acid reduced the final PV by 28.6% and 57.1%, respectively.

Table 4. Peroxide value of margarine prototypes during 49 days of storage (meq O₂ kg⁻¹)

The seed-milk formulations followed the same ranking but showed higher final PV values: 42 meq O₂ kg⁻¹ for the control, 35 meq O₂ kg⁻¹ with essential oil and 18 meq O₂ kg⁻¹ with ascorbic acid. The corresponding reductions were 16.7% and 57.1%. Replacing water with watermelon-seed milk increased the day-49 PV by 20% in the control, 40% in the essential-oil treatment and 20% in the ascorbic-acid treatment.

The lower relative efficacy of basil essential in the seed-milk system (16.7% PV reduction versus 28.6% in the water system) indicates clear matrix dependence. This behavior is compatible with interfacial oxidation mechanisms in which antioxidant effectiveness depends on partitioning and localization at the oil–water interface. The seed-milk phase introduces additional dispersed constituents and may modify antioxidant partitioning and/or promote pro-oxidant interactions; accordingly, efficacy observed in the water system cannot be directly extrapolated to the seed-milk system. [4,26,27]

3.4. Development of Acid Value During Storage

AV also increased progressively (Table 5 and Figure 2B), but the essential oil did not reproduce its PV effect. In the water system, the day-49 AV was 6.5 mg KOH g⁻¹ for the control and 6.8 mg KOH g⁻¹ with essential oil, a 4.6% increase rather than a reduction. In the seed-milk system, essential oil lowered the final AV only from 7.2 to 6.8 mg KOH g⁻¹ (5.6%). Ascorbic acid produced lower final values of 4.5 and 5.0 mg KOH g⁻¹, reductions of 30.8% and 30.6% relative to the corresponding controls.

Table 5. Acid value of margarine prototypes during 49 days of storage (mg KOH g⁻¹)

Figure 2. Peroxide-value (A) and acid-value (B) profiles of water- and watermelon-seed-milk margarine prototypes during storage. Lines connect the observed weekly values; no inferential error bars are shown because independent batch replication was not documented

PV and AV track different deterioration processes: PV reflects primary lipid hydroperoxides, whereas AV reflects free fatty acids generated mainly by hydrolysis. The contrast between the two indices therefore indicates that the basil oil at 0.02% modestly slowed peroxide accumulation but did not consistently inhibit hydrolytic change.

3.5. Performance and Food-Use Considerations

The tested essential oil provided partial protection against primary oxidation, whereas ascorbic acid consistently provided greater protection at the same nominal inclusion level. Oxygenated monoterpenes represented 52.037% of the essential oil, with linalool accounting for 36.079%; their abundance may have contributed to radical-scavenging and/or synergistic effects. However, the present compositional data do not allow the observed antioxidant response to be attributed to any single constituent. [13,14]

The 0.02% essential-oil formulation, combined with the measured estragole proportion of 40.665%, corresponds to a nominal estragole input of approximately 81 mg kg⁻¹ before processing and storage losses. This value is a formulation estimate and should not be interpreted as the residual estragole concentration in the finished product. Because estragole has raised genotoxicity and carcinogenicity concerns, and European evaluations recommend minimizing exposure where practical, further food development should prioritize low-estragole basil chemotypes, quantify residual estragole after processing and storage, and establish product-specific safety before any commercialization claim is made. [28,29]

The present data establish composition and descriptive storage trajectories but do not document independent formulation batches, sensory acceptance, microbiological stability or volatile retention after processing. These endpoints, together with replicated storage trials, are required to define technological relevance and shelf life.

4. Conclusions

Basil leaf essential oil from Brazzaville was characterized as an estragole–linalool chemotype, dominated by estragole (40.665%), linalool (36.079%), and 1,8-cineole (12.757%). At 0.02%, the oil reduced the day-49 peroxide value (PV) by 28.6% in the water-based prototype and by 16.7% in the watermelon-seed-milk prototype, whereas ascorbic acid produced a 57.1% reduction in both matrices. In contrast, the essential oil did not reduce the final acid value (AV) in the water system and lowered it by only 5.6% in the seed-milk system, indicating a greater effect on primary oxidation than on hydrolytic deterioration. The weaker protection observed in the seed-milk matrix suggests that antioxidant efficacy was influenced by matrix-dependent partitioning at the oil–water interface and by possible pro-oxidant interactions with aqueous-phase constituents. These results demonstrate modest, matrix-dependent antioxidant activity at 0.02% and indicate that efficacy observed in one emulsion matrix should not be directly extrapolated to another. Because the essential oil was rich in estragole, further food applications should prioritize low-estragole chemotypes and include verification of residual estragole levels in the finished product before any safety or commercialization claims are made.

ACKNOWLEDGEMENT

The authors acknowledge LEXVA Analytique for chromatographic support and access to its reference database.

Funding

No external funding was reported for this study.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability

All data used in this study are included in the article.