Comprehensive control of quality deterioration and shelf-life extension in sous-vide mussels using a nanoemulsion–liquid smoke hurdle approach

Demet Kocatepe; Bengunur Corapci; Bayram Köstekli; Can Okan Altan; Sonay Başkan; Meryem Yeşim Çelik

doi:10.1039/d6ra07474j

. 2026 Oct 1. Online ahead of print. doi: 10.1039/d6ra07474j

Comprehensive control of quality deterioration and shelf-life extension in sous-vide mussels using a nanoemulsion–liquid smoke hurdle approach

Demet Kocatepe ^a, Bengunur Corapci ^a, Bayram Köstekli ^a,^✉, Can Okan Altan ^a,^b, Sonay Başkan ^b, Meryem Yeşim Çelik ^a

PMCID: PMC13628781 PMID: 42824613

Abstract

The high perishability of farmed Mediterranean mussels (Mytilus galloprovincialis) limits their market distribution, necessitating mild hurdle technologies that preserve freshness without synthetic additives. This study addresses this industrial challenge by evaluating a novel synergistic hurdle technology approach combining sous-vide (SV) thermal processing (85 °C, 10 min), liquid smoke flavorings, and sunflower oil-based nanoemulsions during storage at 3 °C for 27 days. Physicochemical, microbiological, microstructural, molecular, textural, colorimetric, and sensory alterations were thoroughly characterized. Dynamic light scattering (DLS) showed that thermal processing actively altered the nanoemulsion interface, inducing a structural re-organization that significantly reduced droplet size from 253 nm to 190.15 nm, while high physical stability was maintained across groups with zeta potentials ranging from −24.45 ± 0.85 mV to 27.10 ± 3.10 mV. Combining nanoemulsion with liquid smoke (DNS group) established a powerful protective effect against protein and lipid degradation, effectively suppressing total volatile basic nitrogen (TVB-N) well below the regulatory limit of 35 mg N/100 g and minimizing free fatty acid accumulation. Structural integrity modifications characterized via SEM and FTIR revealed a more uniform and cohesive myofibrillar microstructure, likely stabilized by potential phenolic–protein interactions and reduced protein thermal degradation. Crucially, while the control (SV-only) and single nanoemulsion (ANS) groups exceeded microbiological limits by day 18, the combined DNS hurdle approach exerted a superior suppressive effect on microbial proliferation, extending the shelf life up to 24 days (compared to 21 days for liquid smoke alone and 18 days for control) while maintaining higher sensory acceptability. These findings demonstrate that this integrated strategy offers a promising, clean-label method to maximize quality retention in processed bivalves.

A nanoemulsion–liquid smoke hurdle approach significantly inhibits microbial growth, reduces lipid oxidation, and extends the shelf life of sous-vide cooked mussels during refrigerated storage. graphic file with name d6ra07474j-ga.webp

1. Introduction

Seafood serves as a crucial, sustainable protein source for the growing global population. Among various aquatic species, the Mediterranean mussel (Mytilus galloprovincialis) is considered a highly promising food source for the future, owing to its high-quality protein, rich polyunsaturated fatty acid profile, and essential macro and micro element content.¹ The average global consumption of bivalve molluscs stands at 2.6 kg per capita, with global mollusc production reaching 25.8 million tonnes in 2023, thereby accounting for 14% of total fisheries and aquaculture output.² Driven by this expanding demand, Türkiye has recently experienced rapid growth in its aquaculture sector, particularly through mussel farming. Over the past few years, mussel cultivation has emerged as the fourth fastest-growing aquaculture segment in Türkiye, following trout, sea bass, and sea bream.³

Due to their filter-feeding behaviour, mussels are highly susceptible to environmental contamination and exhibit a rapid spoilage rate; however, this product presents a strictly limited shelf life during refrigerated storage, primarily due to microbial deterioration and the potential proliferation of pathogenic bacteria, which ranges from 1 to 2 days and 5 to 6 days.^4–6 Furthermore, non-compliance with hygiene standards during processing can elevate pathogenic microbial loads and accelerate deterioration.⁷ Driven by growing consumer demand for fresh, mildly processed seafood that retains maximum nutritional value, mild thermal treatments have been increasingly adopted. While mild heat-treated and cook-chill products can raise public concerns regarding microbial safety,^8,9 the European Union independently recommends a heat treatment of 60–95 °C for 10–30 min to ensure pathogen-free seafood.¹⁰ Additionally, the high moisture content of mussel meat inherently limits its shelf life, hindering market distribution to non-local regions, which underscores the necessity of processing technologies to extend shelf life and reach new consumer markets.¹¹ However, conventional preservation methods like freezing can impair overall quality by degrading sensory attributes, particularly texture and flavor.¹² This highlights a critical industry gap, necessitating the development of safe, multi-target hurdle preservation strategies to deliver high-value, high-quality farmed mussel products to the market.

To safely extend mussel shelf life and preserve fresh-like sensory attributes, various methods have been employed, including cooking/vacuum packing, smoking, canning, frying/baking, and pasteurization in red sauce.^13–19 Since raw mussel meat spoils rapidly, thermal processing is indispensable, requiring precise temperature–time regulation to optimize quality.¹ In this context, sous-vide (SV) thermal treatment-cooking vacuum-packed food at precisely controlled parameters-minimizes nutrient loss while inactivating degradation enzymes and microorganisms,²⁰ aligning with cook-chill technologies involving cooking below 100 °C, rapid cooling to 0–3 °C, and chilled storage.²¹ For mussels recommend an SV treatment of 85 °C for 10 min;⁸ however, because non-proteolytic Clostridium botulinum strains can produce toxins within 31 days at 3.3 °C, SV-treated seafood shelf life is generally limited to 28 days.²² To overcome this and enhance preservation, pathogen-inactivating liquid smoke fractions can be integrated.²³ Liquid smoke, obtained via biomass pyrolysis, contains carbonyls and phenols that govern product flavor, color, pH, and shelf life through antibacterial and antioxidant activities.^23,24 Offering a safer, standardizable alternative with lower polycyclic aromatic hydrocarbon (PAH) accumulation than traditional smoking, it is applicable via spraying, dipping, or direct incorporation.²⁵

To further optimize the delivery and efficacy of such bioactive antimicrobial treatments in seafood, nanoemulsion systems have emerged as a highly sophisticated approach. These systems effectively enhance antimicrobial activity and facilitate the incorporation of volatile essential oils into food matrices by increasing their distribution in specific zones where microorganisms actively proliferate.²⁶ Recently, the application of nanoemulsions to aquatic products has gained significant momentum. Extant literature reveals that nanoemulsions are primarily utilized in farmed fish species such as trout, sea bream, and sea bass, with studies consistently demonstrating their efficacy in extending the shelf life of these products.^27–31 These investigations have comprehensively evaluated both raw,^28–30,32 and cooked,^29,30 fish models. However, to the best of our knowledge, no literature examines the synergistic effects of combining nanoemulsion, liquid smoke, and SV processing in bivalve molluscs, particularly mussels.

To address the literature gap, this investigation comprehensively evaluated the integration of liquid smoke flavoring and nanoemulsion applications alongside SV thermal processing as a multi-target hurdle technology for mussel preservation. Mussels (6–7 cm) were treated with distinct combinations of these innovative methods and systematically stored at 3 °C for up to 27 days. The preservation efficacy and underlying synergistic impacts were scrutinized through a comprehensive, multi-parametric framework; this included tracking macro-quality indicators such as meat pH, color changes, sensory attributes, total volatile basic nitrogen (TVB-N) accumulation, FFA formation, and microbiological load. Crucially, to provide a deeper understanding of the treatments' effects on the food matrix, tissue structural integrity and molecular characterization were investigated via Scanning Electron Microscopy (SEM) and Fourier-Transform Infrared (FT-IR) spectroscopy.

2. Materials and methods

2.1. Reagents and chemicals

Boric acid (CAS no. 10043-35-3), magnesium oxide (CAS no. 1309-48-4), plate count agar (PCA, 105463.0500), reinforced clostridial agar (RCM, 105411.0500), violet red bile agar (VRBA, 101406.0500), peptone (101800.0500), sodium chloride (106404.1000) (Merck, DE); ethanol (99.5%, CAS no. 9005-65-6, Carlo Erba Regents, FR); Tashiro (Lot no. I260225), phenolphthalein (Lot no. I260225) (ChemBio, Bilge Kimya, TR); liquid paraffin (CAS no. 8012-95-1, QLP, Smart Kimya, TR); hydrochloric acid (Pure, LOT no. O2820, Honeywell Fluka, AT); potassium hydroxide (Pure, CAS no. 1310-58-3, AFG Bioscience, USA); diethyl ether (99.5%, Lot no. LR0563308AGQ, Isolab Chemicals, DE); potato dextrose agar (PDA, ref. no. 4019352, Biolife, IT); oil-soluble liquid smoke (ALF-017, Alfasol, TR); Tween 80 (GRAS, CAS no. 9005-65-6, Sigma-Aldrich, DE). Chemicals from the analytical class were used in the study.

2.2. Mussel material and sample preparation

A total of 77 kg of fresh, live Mediterranean mussels (Mytilus galloprovincialis, Lamarck 1819; shell length: 6–7 cm) were harvested in April 2025 from an offshore longline commercial aquaculture facility located off the coast of Sinop in the Black Sea (class A sanitary classification). Harvesting was performed mechanically using a mechanized mussel harvesting vessel equipped with a continuous lifting and declustering unit to minimize physical trauma and shell breakage. Following harvest, the mussels were packed into food-grade, high-density polyethylene mesh sacks and immediately transferred into insulated containers filled with a refrigerated slurry of pristine offshore seawater and crushed ice (0–2 °C) to prevent desiccation, maintain oxygenation, and suppress baseline metabolic stress. The transport from the offshore harvest platform to the laboratory was completed within approximately 40 minutes under strict cold-chain maintenance. Upon arrival at the laboratory, the mussels were immediately examined; individuals exhibiting cracked shells, gaping valves unresponsive to mechanical stimuli, or signs of mortality were strictly discarded, and viable specimens were gently brushed under chilled potable running water (4 ± 1 °C) to remove epibionts and sediment. Their size and weight were recorded, and flesh yield and condition indices were calculated. For the shell opening process, the mussels were immersed in 100 °C boiling water using a steel sieve for 2 minutes, then removed, placed on trays, and the meat was extracted, and the byssus threads were removed. A total of 9650 g of mussel meat was used in the study. The extracted mussel meat was placed in sieves, excess water was removed, and it was prepared for packaging.

2.3. Preparation of nanoemulsion

The nanoemulsion mixtures were prepared based on the method used by Hamouda et al.³³ Briefly, 2 nanoemulsion formulations were prepared and coded as NA and ND. The compositions were expressed as % (w/w) of the total formulation. NA consisted of 80% distilled water, 14% sunflower oil (Yudum, Savola Food, Türkiye), 3% ethanol, and 3% Tween 80. ND consisted of 80% distilled water, 4% sunflower oil, 3% ethanol, 3% Tween 80, and 10% liquid smoke. The concentration of liquid smoke (resulting in a final concentration of 5% in the mussel treatment mixture) was selected based on preliminary sensory acceptance trials. Higher levels (>5%) imparted an overpowering smoky aroma and an unpleasant bitter aftertaste that completely masked the natural flavor of the mussel meat, whereas 5% provided the most balanced and preferred sensory profile. Oil, ethanol, and Tween 80 were separately added stepwise to each group and stirred using a magnetic stirrer at 900 rpm for a total of 30 min. For the ND formulation, liquid smoke was subsequently added, and stirring was continued for an additional 15 min. The resulting mixtures were then maintained at 86 °C for 1 h and subsequently allowed to cool to room temperature. After cooling, distilled water (80%, w/w) was added to each formulation and homogenized for 15 min (cycle: 7*10%, power: 51%) using probe-type ultrasonicator (HD 2200, Bandelin Sonopuls, Berlin, Germany). The nanoemulsions were prepared and then sterilised under a UV lamp for 20 minutes before being stored at 4 °C.

2.4. Formation of experimental groups

Fresh mussel meat ready for packaging was portioned into 100 g per package prior to group allocation. A total of 8 kg of mussel meat was used for the groups, with 2 kg allocated to each group. The remaining mussel meat was reserved for initial analysis and spare packages. In the nanoemulsion treatment groups (ANS and DNS), the 2 : 1 meat-to-nanoemulsion weight ratio was chosen based on preliminary sensory and handling trials to provide an optimal, uniform surface coating on the mussel meat while preventing excessive free fluid inside the vacuum pouch, which would otherwise result in a soggy texture and diminished sensory appeal after SV cooking. All portioning and packaging procedures were conducted under strictly sanitized and aseptic conditions within a laminar airflow cabinet to prevent cross-contamination. Multi-layer vacuum bags were pre-sterilized under UV radiation for 20 min prior to filling, and packaging was immediately performed using an Abant MG42 (Bolu, Türkiye) vacuum packaging machine at a 95% vacuum level.

For SV processing, PE/PA/PE multilayer vacuum bags (25 × 30 cm; 90 µm thickness) were used. The packaging material had an oxygen permeability of less than 82 cm³ per m² per day and a water vapour permeability of less than 10 g per m² per day, with a heat resistance range of 30–100 °C.

The SV process was carried out using a PackTech IPX7 SV cooker (PackTech Machine, Egypt). The SV treatment was applied at a temperature of 85 °C for a period of 10 minutes. During processing, a thermocouple was inserted into the geometric center of three randomly selected packages from each group to monitor the core temperature. Following SV treatment, the packages were immediately immersed in ice-cold water for 10 min to ensure rapid cooling. Subsequently, all samples were stored at +3 °C for 27 days. Table 1 presents the groups that were used for the experiments.

Table 1. Experimental groups.

Groups	Applied processes
Control	Mussel meat subjected to SV only
SS	Mussel meat treated with 5% oil-soluble liquid smoke and subjected to SV processing
ANS	Mussel meat treated with sunflower oil nanoemulsion and subjected to SV (meat : nanoemulsion ratio 2 : 1)
DNS	Mussel meat treated with sunflower oil + liquid smoke (oil-soluble, 5%) nanoemulsion and subjected to SV (meat : nanoemulsion ratio 2 : 1)

Open in a new tab

2.5. Analyses

Zeta potential and particle size analyses were performed on the nanoemulsions prepared in the study both before and after the SV process prior to the preparation of mussel groups. Following the formation of the experimental mussel groups, at the start of the study, pH, color, TVB-N, free fatty acid content, and microbiological analyses were performed on raw mussel meat, liquid smoke, sunflower oil-containing nanoemulsion, sunflower oil, and liquid smoke-containing nanoemulsions. Sensory analysis, mechanical (texture) analysis, SEM images, and FTIR analysis of raw mussel meat were performed on day-0. From day-0 onwards, pH, color, TVB-N, free fatty acid content, microbiological analyses, sensory and texture analyses were performed every 3 days, including on day-27, for all groups. In addition, SEM imaging analyses were performed on day-0 for all groups formed, and FTIR analyses were performed on both day-0 and at the end of the experiment (day-27). On each sampling day, two independent packages were randomly drawn from each treatment group. Physicochemical and microbiological analyses were conducted in triplicate for each package (n = 6). Mechanical texture profile analyses and color measurements were performed by taking 6 readings on each package (n = 12). Sensory evaluations were assessed by 6 panelists evaluating two different packages (n = 12). All values are presented as mean ± standard error (SE).

2.5.1. Zeta potential (ZPv), mean zeta size (ZSv), and polydispersity index (PDI)

Using dynamic light scattering (DLS) (Malvern Instruments, Worcestershire, United Kingdom), the zeta potential (ZPv), average zeta size (ZSv), and polydispersity index (PDI) of the nanoemulsions (NA: sunflower oil nanoemulsion; ND: nanoemulsion containing liquid smoke; NPD: SV-treated sunflower oil nanoemulsion and NPD: SV-treated nanoemulsion containing liquid smoke) were measured at 25 °C. For every sample, an average of thirty measurements were made.

2.5.2. pH measurement

A digital pH meter (WTW Model ProfiLine 3110, Weilheim, Germany) equipped with a glass electrode was used to monitor the pH levels. Before being measured, homogenized samples were diluted 1 : 10 with purified water.

2.5.3. Colorimetric analyses

A colorimeter (Konica Minolta CR-400, light projection tube type A33a, Tokyo, Japan) based on the CIE L*a*b* color standard was used to assess the samples' color properties.³⁴L* stands for lightness (0 = black, 100 = white), a* for redness (+) or greenness (−), and b* for yellowness (+) or blueness (−) in this system.

2.5.4. Measurement of total volatile basic nitrogen (TVB-N)

The TVB-N process adhered to the Ludorf and Meyer methodology.³⁵ A glass flask containing 2 g of the homogenized material, 5 g of magnesium oxide (MgO), a few drops of liquid paraffin, and one 100 ml of distilled water was moved to the distillation unit. Three to four drops of Tashiro indicator, 5 ml of 3% boric acid, and 50 ml of distilled water were put to an Erlenmeyer flask to collect the distilled liquid. After that, the mixture was brought to boil for 45 min in the distillation unit. After this time, the liquid in the Erlenmeyer flask was titrated with 0.1 N HCl until it turned pink instead of green. Next, the TVB-N value was computed using the pertinent.TVB-N (mg/100 g) = [HCl consumed/sample weight (g)] × 0.14

2.5.5. Free fatty acids (FFAs) analysis

The AOAC technique was used to analyze the free fatty acids.³⁶ A glass vial containing 5 g of the sample was filled with a solution of ethyl alcohol and diethyl ether. Three to four drops of phenolphthalein indicator were then added. After titrating the sample with a 0.5 N KOH solution until a persistent pink color was achieved, the content of free fatty acids was determined by measuring the amount of KOH consumed.% Free fatty acid = [KOH consumed/sample weight (g)] × 2.82

2.5.6. Texture analysis

The mechanical analysis was conducted using a Brookfield CT3 Texture Analyzer (Middleborough, Massachusetts, United States of America). Warner–Bratzler shear (WBS) and texture profile analysis (TPA) analyses were carried out in this situation. A cylindrical probe with a 12 mm diameter and a 5 kg load was installed on the apparatus for the TPA analysis. Analyses were carried out at a probe speed of 2 mm s⁻¹ with the trigger sensitivity set to 0.05 N. Shear tests were conducted with a 50 kg load and the shear force kit was mounted for the WBS analysis. Each mussel had six samples collected at six different times.

2.5.7. SEM (scanning electron microscopy) and FT-IR (Fourier transform infrared spectroscopy) analyses

Mussel samples were dried in a lyophilizer (Telstar, LYOQUEST-85 PLUS ECO, Terrassa, Spain) for SEM and FTIR analysis. The dried mussel samples were analyzed whole for SEM and powdered for F-TIR.

The mussel samples were pre-frozen for 12 hours at −40 °C before being lyophilized in a lab-style lyophilizer. The shelf temperature was gradually raised from −40 °C to −20 °C (0.10–0.20 mbar) during primary drying. Secondary drying was then conducted for a full day at a shelf temperature of +20 °C under a pressure of 0.03–0.05 mbar.

From each dried mussel sample (n = 6), sections of around 1 × 1 cm were produced for scanning electron microscopy (SEM) investigation. A Cressington Sputter Coater 108 Auto (Cressington Scientific Instruments, Watford, United Kingdom) was then used to coat these samples with a 90% Au–10% Pd alloy. A high vacuum FEI Quanta FEG 250 scanning electron microscope (FEI Company, Eindhoven, Netherlands) was used to study the surface morphology of the coated samples. Magnifications of 2500× and 5000× were used to capture the images.

A Bruker Alpha II device (Bruker Optics, Zurich, Switzerland) was used to conduct ATR-FTIR analyses. Each group's 3 mg of lyophilized mussel powder were examined in the 4000–500 cm⁻¹ wavenumber range. 32 scans with a resolution of 4 cm⁻¹ were used to record spectral data. FTIR spectra baseline correction, peak identification and secondary structure-Gaussian deconvolution across the characteristic amide regions (amide A, B, I, II, and III) were processed using OriginPro software (Version 2025b, OriginLab Corporation, Northampton, MA, United States of America) to evaluate conformational modifications in the muscle proteins.

2.5.8. Microbiological analyses

A sterile setting was used for the microbiological sampling procedure. 10 g of the sample were weighed into sterile sample containers after the packaging material's surface was cleaned with 70% ethyl alcohol on each analysis day. After transferring the samples to sterile filter-equipped stomacher bags, 90 ml of peptone was added, and the samples were homogenized in the stomacher for 180 seconds. Sterile 0.85% NaCl solution was used to prepare serial dilutions, and the spread plate method was used for inoculation.

Mesophilic and psychrophilic aerobic bacteria were counted using Plate Count Agar (PCA). Psychrophilic bacteria were cultured for 10 days at 7 °C, whereas mesophilic bacteria were cultured for twenty-four hours at 35 °C. The total number of anaerobic bacteria was measured using Reinforced Clostridial Agar (RCA). After that, Petri dishes were anaerobically incubated in a 10% CO₂ environment for 24 hours at 35–37 °C. Total yeast and mould counts were performed using Potato Dextrose Agar (PDA) in Petri dishes that were cultured at 27–28 °C for 3 to 4 days. Total coliform bacteria counts were determined using Violet Red Bile Agar (VRBA), and the Petri dishes were incubated at 37 °C for a full day. By counting the colonies that grew on the Petri plates, the results were shown as log CFU g⁻¹.

2.5.9. Sensory analysis

Sensory analyses of the samples were conducted using separate sensory analysis forms prepared for the raw mussel and SV groups. A 5-point hedonic scale was used for the mussels (5: liked very much, 4: liked, 3: liked moderately, 2: disliked (edibility threshold value), 1: disliked very much). Samples scoring 2 points or below were considered inedible.

To evaluate raw mussel meat from a sensory perspective such as smell (fresh smell, distinct smell, sea/seaweed smell), appearance of the meat (brightness, wet appearance, orange colour, smooth surface) and texture (firmness (meat tightness), elasticity).

To evaluate cooked mussel meat from a sensory perspective, taste (sour taste, smoky taste, salty taste, sweet taste, umami taste), smell (fresh smell, sea/seaweed smell, smoky smell), texture (firmness, chewiness, juiciness), appearance, and overall evaluation (appearance, fresh smell, sea/seaweed smell, umami taste, firmness, chewiness, and juiciness scores) criteria.

Six trained and chosen evaluators from the Fish Processing Technology Department, comprising postgraduate students and academic personnel, made up the sensory panel. Panelists were chosen based on their lack of seafood allergies, at least five prior panel experiences, and adequate understanding of the standards pertaining to seafood freshness. There was no age categorization or gender balance. The panelists were informed of the study's goals, but it was decided that they did not need any training because of their prior experience. Individual and independent sensory assessments were carried out at room temperature, with panelists not allowed to engage or communicate with one another. Ethical committee approval for the sensory evaluation of cooked mussels was granted by the Sinop University Human Research Ethics Committee (Decision date/no. 05.09.2024/237). Prior to participation on each sampling day, all panelists provided informed written consent by signing a ‘Sensory Analysis Consent Form’, and the evaluation forms were adapted from the methodology described by Oliveira et al.³⁷ The cooked mussel meat's flavor, texture, and scent were rated on a 0–5 point scale. Within each group, two distinct packages were used to draw random samples. The panelists were asked to assess 10 g of mussel samples that were placed on white porcelain plates.

2.5.10. Statistical analysis

Means and standard errors were calculated for all data obtained from the analyses. The Tukey HSD multiple comparison test (p < 0.05) was employed to determine significant differences between means. Additionally, separate one-way analyses of variance (ANOVA) were conducted to identify differences between and within groups across storage days, using OriginPro 2025b software (OriginLab Corporation, Northampton, MA, United States of America). Principal component analysis (PCA) was conducted on seven quality indicators: OA, TVB-N, FFAs, TMAB, TPB, TAN, and TYM. Before analysis, all variables were standardized using z-scores (mean-centred and scaled to unit variance). As a result, PCA used the correlation matrix, ensuring variables with larger numerical scales did not dominate the analysis. The PCA input matrix included 40 observations of group-by-time data (four treatment groups × 10 sampling days), with each observation reflecting the average value for that treatment group and sampling day. PCA was conducted with the prcomp function in R version 4.5.2 within RStudio 2025.09.2, and the outcomes were visualized using ggplot2. PCA was used as an exploratory method to summarize relationships among quality indicators and examine differences between treatment groups and temporal changes in product quality during storage. The biplot and trajectory plot offer complementary visual representations of the same PCA model.

3. Results and discussion

3.1. Biometric characteristics of mussels

In the Black Sea, early spring coincides with the primary reproductive period when Mediterranean mussels attain peak gonadal maturity.³⁸ Consequently, baseline biometric measurements performed on the samples provided important information about the physiological condition and meat quality of the mussels. The mussels had an average shell length of 65.81 ± 0.77 mm, a live weight of 18.22 ± 0.63 g, and a meat weight of 6.86 ± 0.25 g, with a meat yield of 32.13 ± 0.65%. The condition index was 26.00 ± 1.34%, which was found to be high, consistent with the development stage of the gonads. The high condition index indicates that energy reserves are directed particularly towards reproductive tissues.³⁸ The dry matter content in the meat was 18.33 ± 0.36%, reflecting the density of the meat's nutritional content, while the high moisture content (81.67 ± 0.36%) indicates freshness and vitality. The high dry matter content (94.09 ± 0.53%) and low moisture content (5.91 ± 0.53%) in the shell emphasise the mussels' robust external skeleton structure. These data reveal that mussels cultivated in the Black Sea under relatively low salinity, but nutrient-rich conditions can exhibit high quality and productivity during the gonad development period. These findings are consistent with other studies conducted in the Black Sea.^38,39

3.2. Mean zeta size (ZSv), zeta potential (ZPv), and polydispersity index (PDI)

The mean zeta size (ZSv) values of the nanoemulsions prepared in the study before (NA, ND) and after (NPA: NA group treated with SV, NPD: ND group treated with SV) were determined as 230.30 ± 0.80 nm, 253 ± 0.00 nm, 227 ± 20.60 nm and 190.15 ± 24.65 nm, respectively. No statistically significant difference was observed between the nanoemulsion groups (p > 0.05). However, the ZSv value of the ND group is larger than that of NA. The oil-based nature of the liquid smoke may have increased the oil content and thus the density of the nanoemulsion, which could have affected the ZSv size.

The decrease in particle size from 253 nm to 190.15 nm in the ND group following the SV process indicates significant structural changes in the nanoemulsion system. This thermal droplet reduction can be attributed to the combined effects of reduced interfacial tension and enhanced surfactant packing at the oil–water interface during the 85 °C SV treatment. The thermal kinetic energy accelerates the mobility of Tween 80 molecules, while low-molecular-weight amphiphilic phenolic compounds and organic acids present in the liquid smoke partially partition toward the droplet boundary, functioning as co-surfactants that promote spontaneous droplet re-equilibration into a smaller, thermodynamically stabilized colloidal state. Additionally, mild heat treatment disrupts reversible loose droplet flocs or clusters that tend to form in complex pyrolytic oil mixtures at ambient storage. This reduction in droplet size has facilitated more effective contact between the phenolic and organic acid components in the liquid smoke and microbial cells, due to the increase in surface area, thereby enhancing antimicrobial efficacy. The fact that smaller droplets adsorb more easily onto the cell membrane and increase permeability also supports this effect. Furthermore, these observations confirmed that the SV process facilitates the reorganisation of emulsifiers at the interface, thereby creating a more stable and homogeneous structure. Although liquid smoke alone has the effect of increasing droplet size, the formation of smaller and more stable droplets when applied in combination with SV indicates a synergistic effect between the two applications.

The zeta potential (ZPv) value of nanomaterials can be determined as positive or negative mV.³⁰ The zeta potential (ZP) may increase due to the presence of fatty acids in the interface region.⁴⁰ In other words, the carboxyl groups of fatty acids may contribute to a negative charge.^41,42 The ZPv values of the nanoemulsions were measured as −24.45 ± 0.85 mV in the NA group, −23.7 ± 0.00 mV in the ND group, −24.15 ± 1.65 mV in the NPA group after the SV process, and −27.10 ± 3.10 mV in the NPD group. ZPv values above −30 mV indicate a high degree of physical stability.⁴³ In this sense, it can be said that all nanoemulsions in the research groups have high stability values.

The polydispersity index (PDI) indicates the particle size distribution of nano-sized particles. When this value is between 0.1 and 0.25, a narrow distribution close to monodisperse (single distribution) is obtained. When the PDI value is above 0.5, a broad particle size distribution is understood to be present, due to the presence of large-sized particles and aggregate (clump) formation.^44,45 In the study, the PDI values of the groups were determined as 0.386 ± 0.00 in the NA group, 0.422 ± 0.00 in the ND group, 0.310 ± 0.03 in the NPA group, and 0.381 ± 0.12 in the NPD group. The highest PDI value was determined in the nanoemulsion group containing smoke flavouring (ND), while the lowest PDI value was determined in the nanoemulsion group treated with SV (NPA).

The thermal treatment applied in SV was found to have a positive effect on particle distribution in both the NPA and NPD groups. However, it can be said that the particle distribution in groups without liquid smoke was closer to normal distribution. Upon examination of the obtained values, it was determined that the particle size distribution in all groups was within the normally accepted ranges and did not show any signs of agglomeration.

3.3. Color measurement

The color factor is one of the most important elements affecting product quality in foods. Seafood cooking methods, the oil used, the initial quality of the product, the heat treatment temperature and duration, the ingredients added to the product, the cooling conditions applied after the cooking process, and the storage period all affect color. Therefore, in this study, the color parameters (L*, a*, b*) of raw mussel samples, liquid smoke, NA and ND nanoemulsions were measured. When L* values were examined, the lowest brightness value was determined in the raw mussel sample (40.60 ± 0.94) (Fig. 1). The highest L* values were measured in NA (88.54 ± 0.38) and ND (86.29 ± 0.48). The L* value of the smoke flavouring solution (85.59 ± 1.11) was found to be close to these values. These results show that nanoemulsions containing oil have a lighter tone appearance compared to other samples.

The raw mussel sample (4.16 ± 0.25) displayed a positive value based on the a* parameter. The other contents had negative a* values; the liquid smoke (−5.24 ± 0.04), NA (−5.19 ± 0.02), and ND (−5.24 ± 0.04) samples showed comparable unfavorable outcomes. The raw mussel sample (16.73 ± 0.62) had a higher b* value than all other contents when analyzed in terms of b* values. This implies that the yellow tone is formed in part by pigments that are naturally present in mussel meat, such as carotenoids. With NA (10.48 ± 0.05), liquid smoke (9.62 ± 0.39), and ND (9.62 ± 0.39) displaying comparable yellowness values, the b* values of the other constituents were near to one another.

The CIE Lab* system was used to assess the impact of the various treatments used in the study (control, SS, ANS, and DNS groups) on the color parameters (L*, a*, and b*) of mussel meat. At the start (day 0), there were statistically significant variations between the groups' L* levels (p < 0.05). As a result of cooking, the L* value rose. According to other studies, heat treatment raises seafood's brightness value.⁴⁶ However, when applied SV, the anaerobic environment produced by vacuum helps myoglobin stay in the more heat-resistant deoxyhaemoglobin state.⁴⁷ The ANS and DNS groups had the greatest L* values, and there was no statistically significant difference between them (p > 0.05). The control group had the lowest L* value. These results suggest that nanoemulsion and smoke flavoring applications are useful for raising the brightness value of mussel flesh. As the storage period went on, variations in L* values were seen in every group. But for every group, there was no statistically significant difference in L* values between the beginning (day 0) and end of storage (day 27) (p > 0.05).

According to a study on Perna viridis that used SV cooking, samples that were SV cooked had a greater L* value than samples that were raw or regularly boiled.⁴⁸ The temperature, length of time, and type of mussel can all affect the SV method used with them. L* levels were generally high during storage, especially in the ANS group, whereas the control and SS groups showed more noticeable changes. These findings show that applying nanoemulsion increases light reflection from the surface, preserving the brightness value of mussel meat during storage.

When examining the a* (red-green) values, the presence of positive values in all groups on day 0 indicates that the mussel meat has a slightly reddish tone. Initially, the highest a* value was found in the SS (4.71 ± 0.10) group, while the lowest value was found in the ANS (4.07 ± 0.08) group (p < 0.05). However, there was no statistically significant difference in a* values between each group in terms of a* values measured at the beginning (day 0) and at the end of storage (day 27) (p > 0.05). A general upward trend in a* values were observed throughout the storage process. Specifically, the control group reached the highest a* value (7.15 ± 0.37) on day 15. This increase can be attributed to oxidative reactions and the oxidation of myoglobin pigments during the storage process. In contrast, the increase in a* values were more limited in the SS, ANS, and DNS groups. This indicates that the smoke flavouring and antioxidants in the nanoemulsion components slowed down the oxidative color changes. Overall, it can be stated that the addition of nanoemulsion and smoke flavouring had a protective and stabilising effect on the color characteristics of the mussel meat. It was determined that these additives particularly increased the brightness (L*) and yellowness (b*) values and limited color changes during storage. These findings can be explained by the light-reflecting properties of nanoemulsion systems and the antioxidant effect of phenolic compounds in smoke flavouring.

3.4. pH, TVB-N and FFA results

Within the scope of the research, pH measurements were taken for raw mussels, liquid smoke, NA and ND nanoemulsions to enable better interpretation of the pH analysis results, and were found to be 6.55 ± 2.07, 3.13 ± 0.99, 5.57 ± 1.76 and 3.90 ± 1.23, respectively (Fig. 2). Liquid smoke is rich in organic acids formed during the pyrolysis process and liquid smoke itself generally has a very low pH between 2.8 and 3.1.⁴⁹ In this study, the pH value of liquid smoke was determined to be 3.13. The applied thermal treatment had a greater effect on pH, and the pH value, which was initially 6.55 with the SV process, increased in all groups. The initial pH values of the groups subjected to SV were similar to each other (p > 0.05). The thermal treatment applied to foods affects pH. One study emphasised that the pH value of sea bass increased with the SV cooking process.⁵⁰ This can be attributed to the formation of disulfide formation due to the alkaline environment created during the cooking process. Higher pH values cause an increase in disulphide formation.⁵¹ The liquid smoke application did not cause a significant difference in the pH values of the SS and DNS groups during the first days of storage but became more effective towards the end of storage. The SS and ANS groups were more stable in terms of pH during storage, with the ANS group having relatively stable and high pH values throughout storage. Since the pH value of the nanoemulsion used in the ANS group was 5.57, the pH of the product was higher from the outset. In the DNS group, the pH value tended to decrease during storage compared to day 0. The very low pH value (3.13) of the liquid smoke triggered a decrease in pH in these groups throughout storage.

The liquid smoke applied to the SS and DNS groups strengthened the acidic structure of the product due to phenolic compounds and organic acids. In other words, as it contained both the natural acidic structure of the liquid smoke and the carrier effect of the nanoemulsion, it had a lower pH value during long-term storage compared to the other groups.

TVB-N is directly related to protein degradation, which usually occurs because of microbial activity and enzymatic breakdown. In order to better interpret the TVB-N analysis results within the scope of the study, TVB-N measurements were also performed on raw mussels, liquid smoke, NA and ND nanoemulsions, and these data were found to be 17.18 ± 0.23, 5.12 ± 0.21, 2.21 ± 0.11 and 3.28 ± 0.16 mg N/100 g, respectively (Fig. 2).

The findings regarding the TVB-N values for the groups are given in Fig. 2. The TVB-N value of NA is quite low. From this perspective, it can be said that the nanoemulsion may have a protective effect against protein degradation due to its pure oil structure and the stability of the nanoemulsion. The combination of liquid smoke and nanoemulsion also had an effect on the decrease in the TVB-N value of raw mussels. The initial TVB-N value in raw mussel meat was found to be relatively high. However, after SV application, this value decreased significantly in all groups, and lower values were obtained. The temperature applied in the SV process affects the TVB-N value. Gundermann reported that the increase in TVB-N in mussels cooked SV at 72 °C was faster than in mussels cooked SV at 80 °C and above.⁵² This indicates that the SV cooking temperature is effective in reducing microbial activity and limiting protein degradation at the outset. Furthermore, in the control group, TVB-N values rose rapidly as storage progressed, reaching 18.59 mg N/100 g on day 3 (p < 0.05). This finding reveals that SV processing alone does not provide sufficient protection for long-term storage. The additional applications had significant effects on TVB-N formation. In the SS group, although the liquid smoke application (pH 3.13) initially delayed the increase, a rapid rise in TVB-N values was observed from the 12th day onwards, reaching the highest levels on days 15–18 (p > 0.05). Researchers have reported that the application of nanoemulsions to different aquatic products suppresses the increase in TVB-N.^30,53 In the ANS group (NA pH: 5.57), TVB-N values remained relatively stable during storage (except on day 9) and were significantly lower compared to the control and SS groups (p < 0.05). In the DNS group (ND pH: 3.90), a marked decrease in TVB-N values was observed during the initial stages of storage. This indicates that the synergistic effect of nanoemulsion and liquid smoke limited both microbial growth and the formation of volatile nitrogen compounds associated with protein degradation. Although partial increases occurred in the following days, TVB-N levels in the DNS group remained lower than in the control and SS groups. This finding demonstrates the potential of combining nanoemulsions and liquid smoke to delay protein degradation and increase the product's resistance to spoilage. Throughout the storage period, the TVB-N content of all groups did not exceed the limit value of 35 mg N/100 g specified by European Community Directive 95/149.⁵⁴

Free fatty acid measurements were performed on raw mussels, liquid smoke, NA and ND nanoemulsions within the scope of the study, and these data were found to be 6.13 ± 0.55, 1.74 ± 0.23, 0.94 ± 0.20 and 0.61 ± 0.07 mg/100 g, respectively (Fig. 2).

Unsaturated fatty acids can undergo oxidation after a complex series of reactions in the presence of free radicals and oxygen. The SV process is effective in reducing lipid oxidation in foods.⁵² Free fatty acid determination is a marker used to detect oxidation in fatty foods. To date, there is no study in the literature investigating the synergistic effect of SV + liquid smoke + nanoemulsion on FFAs. The SV treatment applied to mussel meat reduced the FFA values of the groups. However, an increase was observed in all groups during the storage period. Corapci et al. investigated the effect of lemon oil-containing nanoemulsion on cooked and raw fish meat and reported that the FFA content of the groups decreased on the first day of the study and increased during storage, similar to this study.²⁹ The main difference between the two studies is that Corapci et al.²⁹ did not apply heat treatment to the nanoemulsion but treated cooked fish meat with the nanoemulsion. In the present study, the heat treatment applied to nanoemulsions caused an increase in FFAs.

The lowest free fatty acid values on day 0 of storage were detected in the ANS and DNS groups. The SV application, together with the nanoemulsion application, reduced FFAs. However, fluctuations in FFA values were observed in all groups during storage. Except for days 18 and 27 of storage, the lowest FFA values were detected in the ANS group. On all storage days except days 3, 15, 24, and 27, the DNS group exhibited lower FFA values than the ANS group. The main reason for the higher FFA values in the SS and DNS groups is that liquid smoke is soluble in lipid and has high acidity. In the ANS group, it can be said that the SV application and the nanoemulsion additive had a positive effect on the FFA values of the mussel samples. For food-grade fish lipids, it is recommended that free fatty acids be between 2% and 5%.⁵⁵ The FFA content of the groups did not exceed the limit values during storage.

3.5. Texture analysis results

The texture profile analysis (TPA) parameters of the fresh mussel meat used in the study as follows; hardness, adhesiveness, resilience, springiness, cohesiveness and chewiness, were determined as 4.75 ± 0.11 N, 0.19 ± 0.03 mJ, 0.07 ± 0.01, 2.52 ± 0.15 mm, 0.13 ± 0.01 and 1.57 ± 0.13 N mm, respectively (Fig. 3). For fresh New Zealand mussels (Perna canaliculus), hardness, cohesiveness, springiness and chewiness values of 4.99 ± 0.53 N, 0.32 ± 0.04, 3.70 ± 0.90 and 5.95 ± 2.05 N mm were reported by Xin et al.⁵ No TPA data for fresh Mediterranean mussel meat (M. galloprovincialis) could be found in the literature. However, a hardness value of 1.57 N has been reported by Gu for fresh mussels of this species,⁵⁶ which is approximately one-third of the hardness value measured in our study. Hardness is closely related to the size and weight of the mussels, but these parameters were not given in that study. The geometry of the probe and the test settings are further sources of difference between studies. For sous-vide cooked M. galloprovincialis stored at 3 °C, Bongiorno et al. reported that the turgidity of the meat, assessed by sensory evaluation, was better maintained when the mussels were packed with brine, and related this to the higher moisture content of the brined product (81.4% against 73.5%).⁸ In the present study, the nanoemulsion groups were packed with a liquid phase in a comparable way, and their lower hardness on day-0 may in part reflect a similar moisture effect in addition to the lipid–protein interactions discussed below.

On day-0, the highest hardness value was measured in the control group (14.96 ± 0.16 N), followed by the SS, ANS and DNS groups. The differences between the groups on day-0 were not statistically significant (p > 0.05), but the lower values of the ANS and DNS groups were consistent. An increase in hardness would normally be expected in the groups treated with liquid smoke, since the carbonyl compounds of smoke react with muscle proteins and firm the surface of smoked products.⁵⁷ Such an increase was not observed. Whether liquid smoke firms the tissue depends on the composition of the preparation. In salmon fillets treated with two commercial liquid smoke flavourings and stored at 2 °C for 45 days, one flavouring left hardness, gumminess and chewiness at the level of the untreated control, whereas the other raised them and also reduced the changes caused by storage.⁵⁸ Changes in the TPA parameters of meat products by liquid smoke flavourings have likewise been shown to differ between preparations.⁵⁹ The oil-soluble flavouring used here, applied at 5%, evidently belongs to the first type. In the SS group the hardness was slightly below that of the control, and in the DNS group, where the smoke was applied within the nanoemulsion, the hardness was lower still. It is therefore considered that the effect of the nanoemulsion on the muscle tissue outweighed any firming effect of the smoke carbonyls. The lower hardness of the ANS and DNS groups from day-0 to the end of storage indicates that the nanoemulsion application softened the muscle tissue from the outset. It has been reported that reducing the size of the lipid droplets in emulsions applied to protein-based foods modifies texture by increasing the lipid–protein interactions.⁶⁰ Nano-sized droplets can penetrate the muscle matrix more effectively than larger droplets, and the myofibrillar structure is relaxed as a result; the uniform porous structure of the ANS group in the SEM images and the lipid bands in the FT-IR spectra (Fig. 5) are consistent with this explanation. A decrease in hardness was observed in all groups during storage. This decrease is attributed to proteolytic activity and to the continuation of the structural changes that had been initiated by the partial denaturation of the myofibrillar proteins during SV cooking. Softening of fish muscle during chilled storage has been related to the activity of endogenous proteolytic enzymes, in particular cathepsins and collagenases, which degrade myofibrillar and connective tissue proteins,⁶¹ and the rate of softening has been shown to increase with storage temperature.⁶² Since SV cooking at 85 °C is likely to inactivate only part of this enzyme activity, residual proteolysis together with bacterial enzymes is considered responsible for the gradual loss of hardness, cohesiveness and chewiness observed here. On day-0, the adhesiveness of the ANS and DNS groups was significantly lower than that of the two groups without nanoemulsion. A similar initial reduction of adhesiveness by nanoemulsion treatment was reported for fish meat by Chuesiang et al.⁵³ During storage, adhesiveness increased in all groups. The increase was most pronounced in the ANS group on days 24 and 27 (0.42 and 0.50 mJ, respectively), where the values were significantly higher than those of the other groups (control: 0.36 mJ; DNS: 0.30 mJ; SS: 0.40 mJ). The high adhesiveness of the ANS group is attributed to changes in the surface properties of the tissue caused by the lipid–protein complexes formed after the homogeneous penetration of the nanoemulsion. The formation of a film layer on the surface and the release of a larger proportion of myofibrillar proteins are known to increase adhesiveness. The accumulation of free fatty acids and protein degradation products on the surface during storage may also have contributed. The marked increase in the lipid/protein intensity ratio in the day-27 FT-IR spectrum of the ANS group (Fig. 5) supports the presence of oil at the surface of these samples.

Resilience values ranged between 0.188 and 0.228 in all groups, and no significant differences were found between the groups (p > 0.05). The elastic recovery of the mussel muscle tissue was therefore largely unaffected by the treatments.

On day-0, the highest springiness values were measured in the control (5.81 mm) and DNS (5.75 mm) groups, and lower values in the ANS (5.24 mm) and SS (4.93 mm) groups. By day-27, springiness had been best preserved in the ANS group (12% loss), whereas the largest loss (36%) was recorded in the DNS group. Phenolic and carbonyl compounds are known to interact with myofibrillar proteins and, under suitable conditions, to form covalent cross-links.^63,64 In the present study, however, no spectral evidence of covalent cross-linking was found at the smoke level used, and the interactions between the smoke components and the proteins are considered to be mainly non-covalent (Fig. 5). Such interactions appear to have stabilised the myofibrillar network at the beginning of storage, as reflected in the high initial springiness of the DNS group, and to have loosened during storage, which would explain the rapid loss of springiness in this group. The better preservation of springiness in the ANS group indicates that the nanoemulsion without liquid smoke modified the protein structure to a lesser extent, so that softening and loss of springiness were delayed.

Cohesiveness decreased during storage in all groups, and the control group showed the largest fluctuations. On day-0, the highest cohesiveness values were measured in the SS (0.515 ± 0.02) and control (0.503 ± 0.01) groups and lower values in the ANS (0.485 ± 0.02) and DNS (0.477 ± 0.03) groups, although the differences were not statistically significant. The lower initial values of the nanoemulsion groups suggest that the internal bond strength of the muscle tissue was reduced by the nanoemulsion application from the outset. By day-27, the values of all groups had converged to 0.423–0.454, which indicates that proteolysis had reached a similar level in all groups. A decrease in cohesiveness is generally associated with the breakdown of myofibrillar proteins and the deterioration of the muscle fibre structure. On day-0, the highest chewiness values were measured in the control (28.86 ± 0.76 N mm) and SS (27.98 ± 1.02 N mm) groups, and slightly lower values in the ANS (24.13 ± 1.05 N mm) and DNS (24.39 ± 0.45 N mm) groups. At the end of storage (day-27), the lowest chewiness was found in the DNS group (12.98 ± 0.44 N mm, 47% loss), followed by the ANS (14.55 ± 0.37 N mm, 40% loss), SS (17.38 ± 0.43 N mm, 38% loss) and control (24.04 ± 0.39 N mm, 17% loss) groups. The larger chewiness losses of the nanoemulsion groups indicate that the protein matrix was modified by these treatments in a way that reduced its textural stability during storage. Partial protein denaturation during SV cooking, combined with the penetration of the nanoemulsion, is likely to have weakened the structural integrity of the muscle tissue. The largest loss, in the DNS group, shows that the smoke components carried by the nanoemulsion added to this effect.

Taken together, the TPA results show that the textural properties of the mussel meat were modified to a considerable extent by the nanoemulsion application. Springiness was best preserved in the ANS group, but this group also showed the highest adhesiveness and a large loss of chewiness. In the DNS group, the addition of liquid smoke to the nanoemulsion led to low initial hardness and to faster textural losses during storage. This is attributed to the phenolic and carbonyl compounds of the liquid smoke, which appear to have affected the stability of the lipid–protein and phenol–protein associations within the tissue rather than the secondary structure of the proteins themselves (Fig. 5).

3.6. SEM and FT-IR analyses results

SEM and FT-IR provide information about product quality by utilising critical changes such as protein denaturation, structural damage, and molecular interactions that occur during the processing of aquatic products at molecular and microscopic levels. Therefore, SEM and FT-IR are complementary analyses. While SEM shows physical damage, FT-IR is concerned with the chemical mechanism behind this damage. Within the scope of the study, SEM and FT-IR imaging analysis were performed on raw mussels at the beginning of the experiment and on all groups at the beginning of storage and on day 27. SEM analysis images are provided in Fig. 4.

Untreated fresh mussel meat (Fig. 4a) exhibited a smooth and compact surface in which the myofibrils formed a continuous plane. It is believed that most of the observed discontinuities are narrow cracks formed by the removal of water from the tissue via sublimation during freeze-drying. In the SV-cooked control samples on day-0 (Fig. 4b), the surface morphology was markedly different. It had acquired a three-dimensional character, with lamellar protein sheets and deep voids between the myofibrils. These changes can be regarded as the microstructural counterpart of the thermal denaturation of myosin and actin, which unfold and re-associate during cooking and thereby loosen and rearrange the fibre structure.⁶⁵ A similar sequence has been reported for sous-vide tilapia, where the bonds between muscle fibres were broken, the sarcolemma was fragmented and the tissue became more brittle, and where the interstitial spaces became denser and more compact when the cooking temperature was raised to 70–80 °C.⁶⁶ The lamellar sheets seen in the images are the freeze-dried form of the denatured protein aggregates. After 27 days of storage (Fig. 4c), the surface of the control group had become more heterogeneous, and signs of fragmentation were observed in the protein matrix. These changes are attributed to the enzymatic and autolytic activity that continued during storage.

In the ANS group on day-0 (Fig. 4d), a convoluted and wavy surface with a uniform porous structure was observed, which was not present in any other group. Nano-sized lipid droplets are known to penetrate the muscle matrix more effectively than the droplets of a conventional emulsion. The pores are voids left behind by the sublimation of droplets homogeneously distributed among the myofibrils; the FT-IR results discussed in Fig. 5 also support our findings regarding the presence of lipid within the tissue. After 27 days (Fig. 4e), the pores had widened and macroscopic voids had formed in places, but the convoluted surface morphology was largely preserved, and the structural integrity was better maintained than in the other groups.

The DNS group on day-0 (Fig. 4f) showed the same curvilinear basic structure as the ANS group, together with a granular and rough texture in some areas. This granular texture is mainly attributed to the interaction of the phenolic components of the liquid smoke, carried into the tissue within the nanoemulsion droplets, with the muscle proteins. Guaiacol, syringol and 4-methylguaiacol are the main volatile phenols of commercial wood-smoke condensates, their proportions depending on the botanical origin and the pyrolysis conditions. These compounds can bind to proteins through hydrogen bonding and hydrophobic association, and the carbonyl compounds of the smoke can react with free amino groups.^57,64,67 After 27 days (Fig. 4g), the surface of the DNS group had become heterogeneous. In some areas the convoluted porous structure was preserved, whereas in others the surface was compact and flattened. It is possible that the smoke components were not distributed evenly within the tissue, or that the complexes formed with the proteins were not equally stable in all regions.

The SS group on day-0 (Fig. 4h) differed from both nanoemulsion groups. Smooth and compact regions resembling the raw meat alternated with convoluted and processed regions. The conventional liquid smoke application contains macro-sized droplets and reaches only the surface layer of the tissue; where the smoke reached the tissue it was modified, and where it did not the original structure was retained. After 27 days (Fig. 4i), this contrast had become sharper, with flat compact islands surrounded by wavy edges. It appears that a protective layer was formed by the liquid smoke at the surface,⁶⁷ whereas in the deeper tissue the effects of SV-induced denaturation were dominant.

When SEM and TPA results are evaluated together, it becomes clearer to observe the changes in the groups. The compact and continuous myofibrillar structure of the control group corresponds to its high hardness and chewiness values. The uniform porous structure of the ANS group corresponds to the loosening of the myofibrillar network by the lipid droplets and to the lower hardness and chewiness of this group. In the DNS group, the granular and, later, heterogeneous surface is consistent with mostly non-covalent (hydrogen bonding and hydrophobic) and weak complexes formed between the phenolic and carbonyl compounds of the liquid smoke and the myofibrillar proteins. These complexes appear to have stabilised the network at the beginning of storage and to have loosened later, which, together with proteolysis, accounts for the larger losses of elasticity and chewiness observed in this group. The disorganisation, void formation and fragmentation seen in the micrographs during storage follow the same trend as the decreases in hardness, cohesiveness and chewiness in the TPA measurements.

The spectra of the raw meat and of the four groups on day-0 and day-27 are shown in Fig. 5a and b, and the amide I region is examined in detail in Fig. 5c–e. The five amide bands of muscle proteins were present in every spectrum: amide A at 3264–3276 cm⁻¹, amide B near 3070 cm⁻¹, amide I at 1628–1640 cm⁻¹, amide II at 1521–1537 cm⁻¹ and amide III at 1225–1233 cm⁻¹.^68–70 Lipid bands were observed at 2923 and 2853 cm⁻¹ (CH₂ stretching), at 3011–3014 cm⁻¹ (cis = C–H stretching) and at 1741–1745 cm⁻¹ (triglyceride ester C O stretching),^71,72 and the C–O stretching band of glycogen and other carbohydrates, a prominent feature of Mytilus tissue,⁷³ was observed at 1022–1034 cm⁻¹. Since ATR intensities depend on the contact between the powder and the crystal, band positions, band shapes and intensity ratios were compared rather than absolute absorbances. It should also be noted that the raw sample had been blanched for two minutes to open the shells and is therefore not a native tissue.

The protein bands were changed by SV cooking in the same way in all four groups. The amide I maximum was shifted from 1640 cm⁻¹ in the raw meat to 1628–1631 cm⁻¹ in the cooked groups, and the bandwidth at half height increased from 54 to 61–64 cm⁻¹ (Fig. 5c). In the second-derivative spectra (Fig. 5d), a strong component at 1623 cm⁻¹ and a weak component at 1691 cm⁻¹ were observed in all cooked groups; neither was present in the raw meat. By curve fitting of the amide I band (Fig. 5e), the α-helix content was estimated as 27.0% in the raw meat and 19–24% in the cooked groups, and the β-sheet content as 36.7% and 41–44%, respectively. Components at 1620–1625 and 1690 cm⁻¹ are assigned to intermolecular β-sheets, the structures formed when unfolded protein chains associate with one another.^70,74 The same α-helix to β-sheet transition has been followed by FT-IR microspectroscopy in heated beef and pork,^75,76 and has been reported for sous-vide tilapia, where it was accompanied by wider gaps between the fibres in SEM and by a change in hardness.⁶⁶ The same pairing was observed in the present study (Fig. 3 and 4b). The amide A maximum was shifted from 3264 to 3267–3271 cm⁻¹ and the amide II maximum from 1535 to 1521–1529 cm⁻¹, which indicates weaker N–H⋯O C hydrogen bonding along the peptide backbone after cooking; a free N–H group absorbs at 3400–3440 cm⁻¹, whereas a hydrogen-bonded N–H group absorbs near 3300 cm⁻¹.⁷⁷ The amide III band, in which C–N stretching and N–H bending is combined with C O in-plane bending and C–C stretching,⁶⁹ was shifted from 1225 to 1228–1232 cm⁻¹.

On day-0, no difference between the treatments and the control could be detected in the amide I band. The maximum, the width and the fitted composition were nearly the same in all four groups (Fig. 5c–e); the only exception was the SS group, in which the α-helix content (18.9%) was lower than in the control (23.6%). Differences between the groups were found instead in the amide II band, which was located at 1524 cm⁻¹ in SS, 1527 cm⁻¹ in DNS and 1529 cm⁻¹ in ANS, against 1521 cm⁻¹ in the control. A comparable division was reported for smoked salmon by Carton et al.: the amide I band was affected by salting and heat, whereas the amide II band was affected by smoking, which was explained by the reaction of smoke carbonyls with amino-acid side chains, whose contribution to amide II is larger than to amide I.⁷⁸ In the present study the amide II band was also shifted in the ANS group, so the shift cannot be regarded as specific to smoke. It is interpreted instead as a general indication that the added components interacted with the side chains and the N–H environment of the proteins rather than with the backbone conformation. In the SS and DNS groups these interactions are expected to be hydrogen bonding, hydrophobic association and carbonyl–amine reactions of the smoke phenols and carbonyls,^57,64,67,79 and in the ANS and DNS groups the hydrophobic association of the sunflower-oil droplets with the proteins.⁶⁰ The type of bond that dominates depends on the phenol; for sea bass myofibrillar protein, gallic acid has been reported to bind mainly through hydrophobic forces, caffeic acid through electrostatic forces, and catechin or chlorogenic acid through hydrogen bonds and van der Waals forces.⁸⁰

Raised from the SEM images, two observations are relevant to the hypothesis that the smoke phenols interact with the protein matrix. First, the α-helix content on day-0 was lowered only by the non-encapsulated smoke (SS); with the encapsulated smoke (DNS) the secondary structure could not be distinguished from that of the control. The access of the phenols and carbonyls to the protein backbone therefore appears to be limited by the nanoemulsion droplets, while side-chain interactions are still permitted, as shown by the amide II shift in the DNS group. Second, no additional amide I component that could be assigned to covalent cross-links was resolved at the 5% liquid smoke level. The interactions are therefore considered to be mainly non-covalent, in agreement with the small differences in hardness and springiness between the DNS and control groups on day-0 (Fig. 3). The independent evidence that the nanoemulsion had entered the tissue is provided by the lipid bands: on day-0 the ester C O/amide I intensity ratio was 0.35 in ANS and 0.37 in DNS, against 0.23 in the control and 0.25 in SS. The higher value of the control compared with the raw meat (0.07) is probably a concentration effect, since water-soluble sarcoplasmic proteins are lost with the drip during cooking.

The spectra were changed by storage in a group-dependent manner. In the control and ANS groups, the α-helix content decreased from 23.6 to 16.0% and from 23.0 to 12.9%, respectively, while the β-turn and aggregate components increased from 22.9 to 27.3% and from 23.5 to 28.7% (Fig. 5e). In the SS and DNS groups, the α-helix content was hardly changed (18.9 to 20.6% and 23.8 to 19.9%). A loss of α-helix during chilled storage, accompanied by an increase in TVB-N and by fragmentation of the myofibrils, has been reported for mackerel and is the expected result of endogenous and bacterial proteolysis;⁸¹ it has also been shown that the total bacterial count of stored mussels can be predicted from FT-IR spectra with an R² of 0.89.⁸² The retention of ordered structure in the two smoke-containing groups is in agreement with the stabilising effect of moderate amounts of phenolic compounds on myofibrillar protein gels, in which a dose of about 20 mg g⁻¹ was found to raise the α-helix content by about 10%,⁸³ and, for the DNS group, with the lowest TVB-N and microbial counts of the study (Fig. 2 and 6). The picture is not uniform, however. Between day-0 and day-27, the amide A band was shifted upward by 4–7 cm⁻¹ in every group, and the amide II band of the three treated groups was shifted back from 1524–1529 cm⁻¹ to 1532–1537 cm⁻¹, whereas that of the control remained at 1521–1522 cm⁻¹. It can therefore be stated that the hydrogen bonds were weakened in all groups and that the side-chain interactions established by the additives on day-0 were rearranged during storage.

The DNS group showed the best chemical, microbiological and secondary-structure stability of the four groups and, at the same time, the largest losses of elasticity (36% of springiness) and chewiness (47%). Since the protein backbone was preserved, the texture losses of the DNS group cannot be attributed to unfolding of the proteins. They are attributed instead to the relaxation of the non-covalent phenol–protein and lipid–protein associations that had stabilised the network on day-0, which is also suggested by the alternating convoluted and flattened regions of the day-27 micrograph (Fig. 4g). In the ANS group, which contains no liquid smoke phenols, the loss of α-helix was accompanied by a large increase in the lipid/protein intensity ratio (ester C O/amide I from 0.35 to 1.36). Migration of oil to the surface of the tissue during storage is a possible explanation, and it would be consistent with the sharp increase in the adhesiveness of the ANS samples on days 24–27 and with the surface film discussed before (Fig. 3); since a single spectrum was recorded per group, a contribution from sampling heterogeneity cannot be excluded. In the DNS group the ratio remained at 0.33–0.37, which indicates that the oil was retained within the matrix. No free fatty acid C O shoulder near 1710 cm⁻¹, no trans C C band near 968 cm⁻¹ and no consistent decrease of the cis = C–H/CH₂ ratio (0.18–0.25 in all spectra) were detected, in agreement with the low FFA values recorded throughout storage (Fig. 2); the extent of lipid hydrolysis and oxidation therefore remained below the detection limit of the spectra.⁷¹ The carbohydrate C–O band decreased relative to amide I in the two smoke-containing groups (from 0.28 to 0.23 in SS and from 0.33 to 0.23 in DNS) and was unchanged in the control (0.26), which is taken as an indication of glycogen consumption by post-mortem and microbial metabolism.

3.7. Microbiological analysis

All groups underwent periodic microbiological investigations throughout during the storage period. Additionally, raw mussel meat, liquid smoke, NA, and ND nanoemulsions were subjected to enumeration of total mesophilic aerobic bacteria (TMAB), total psychrophilic aerobic bacteria (TPB), total anaerobic bacteria (TAN), total yeast and mould (TYM), and total coliform bacteria (TC). Raw mussels, liquid smoke, NA, and ND nanoemulsions had TMAB and TPB counts of 2.96 ± 0.01 and 3.36 ± 0.04, 1.20 ± 0.19 and 0.85 ± 0.15, 1.57 ± 0.03 and 0.70 ± 0.00, and 1.63 ± 0.03 and 0.85 ± 0.15 log CFU g⁻¹, respectively (Fig. 6). Raw mussel meat, liquid smoke, NA, and ND exhibited TAN, TC, and TYM loads of less than 0.30 log CFU g⁻¹. The mussels used in this study were harvested from a class A aquaculture area subject to continuous sanitary monitoring and devoid of baseline microbiological risks.

All the SV groups' TC counts during storage were found to be less than 0.30 log CFU g⁻¹. As the storage period went on, all groups generally saw an increase in TMAB, TPB, TAN, and TYM loads; however, the degree of this rise differed based on the application groups. The literature does not provide any specific study data regarding the synergistic effect of nanoemulsion + liquid smoke + SV treatment on mussel meat. TMAB values, which were comparable in all groups at the start of storage, dramatically increased over time, especially in the control group. The control group, which only received SV, reached a level of 5.24 log CFU g⁻¹ on day 18 and exceeded the limit value (5.0 log CFU g⁻¹). Therefore, the microbiological shelf life of the control group can be estimated to be approximately 18 days. Therefore, the microbiological shelf life of the control group can be estimated to be approximately 18 days. The TMAB value of the SS group was 4.67 log CFU g⁻¹ on day 18 and reached 5.36 log CFU g⁻¹ on day 21. In this case, the shelf life of the SS group is approximately 21 days. Liquid smoke extended the shelf life by approximately 3 days compared to the control group. On day 18, the ANS group's TMAB value was 5.18 log CFU g⁻¹. This number is just over the upper limit. As a result, the ANS group's shelf life can be calculated to be roughly 18 days. On day 18, the DNS group's TMAB value was 3.86 log CFU g⁻¹, which is significantly below the limit. On day 21, it was 4.40 log CFU g⁻¹, and on day 24, it was 5.64 log CFU g⁻¹. These findings indicate that on day 24, the DNS group's TMAB value was higher than the 5.0 log CFU g⁻¹ limit. Consequently, the microbiological shelf life of the DNS group can be estimated to be approximately 24 days. The synergistic impact of liquid smoke components (phenols, carbonyls, etc.) with the nanoemulsion may be the cause of the notable difference seen in the DNS group. The microbial suppression impact was restricted because the nanoemulsion application by itself lacks this synergy. Both the direct antibacterial impact of liquid smoke and the stabilization of active ingredients in the lipid phase may have improved the nanoemulsion. Contrary to expectations, the ANS group's shelf life was comparable to the control group's; this suggests that the nanoemulsion's antimicrobial efficacy may have declined under oxygen-free or heat-processed conditions, but when combined with smoke flavoring (DNS), it produced a synergistic antimicrobial effect that greatly extended the shelf life.

The total psychrophilic aerobic bacteria (TPB) load of the groups were statistically similar in the first days of storage (days 0, 3, and 6) (p > 0.05). This similarity continued in the control, SS, and ANS groups on days 15, 18, and 21 of storage (p > 0.05). During storage, the increase was more pronounced in the control group and more controlled in the DNS and ANS groups. This indicates that the nanoemulsion has an inhibitory effect on microorganisms that develop at low temperatures.

The possibility of dangerous bacteria producing anaerobic spores during heat treatment is the SV process's biggest drawback.^84,85 Because psychotropic Clostridium botulinum strains are more common in fish and other seafood,⁸⁶ seafood cooked with SV may carry a higher risk of botulism than other food products processed in this manner. After SV processing, goods should be quickly chilled to <4 °C to avoid this.⁸⁷ SV-processed seafood should be cooked before consumption and kept between 0 and 3 °C to avoid chemical and microbiological deterioration.⁸⁸

The TAN levels of the SV groups were <0.30 log CFU g⁻¹ in all groups during the first days of storage (p > 0.05). From the ninth day onwards, the TAN load increased. However, while this increase was higher in the control and SS groups, it remained low in the ANS and especially the DNS groups. This result indicates that the nanoemulsion is also partially effective against microorganisms that develop under oxygen-free conditions. Even on the last day of storage, the TAN bacterial load did not exceed 5.50 log CFU g⁻¹ in all groups. There are no studies in the literature investigating the effect of SV cooking on the microbial load of products treated with nanoemulsion. In their study, Corapci et al.³⁰ treated fish meat with a nanoemulsion solution after cooking and examined its shelf life. Similar to the present study, the TAN load of the groups did not exceed 5 log CFU g⁻¹ in this study.

The TYM load of raw mussel meat is 2.55 ± 0.08 log CFU g⁻¹. With the SV application, the TYM load decreased to <0.30 log CFU g⁻¹. While a significant increase was observed in the control group during the later stages of storage, it was noted that the increase was quite limited in the DNS and ANS groups. The microbiological quality of mussels processed using the SV method is directly related to the cooking temperature and duration applied. Increased temperature significantly affects microbiological stability. Russo et al. reported that mussels cooked at temperatures above 80 °C (90 °C and 100 °C) yielded much better results microbiologically than those cooked at 72 °C.¹⁴ Temperatures of 80 °C and above inhibit the growth of Pseudomonas species, yeasts, and moulds.

When comparing the microbiological shelf lives of the groups: the control group (SV only) had a shelf life of 18 days, the ANS group (nanoemulsion only) had a shelf life of 18 days, the SS group (liquid smoke only) had a shelf life of 21 days, while the DNS group (liquid smoke + nanoemulsion combination) reached a shelf life of 24 days. It was observed that nanoemulsion (ANS) alone did not provide any additional contribution to shelf life, while liquid smoke only extended it by 3 days; however, the combination of the two (DNS) provided the highest protection by extending shelf life by 6 days compared to the control group. SV mussels containing preservatives (citric acid, potassium benzoate, etc.) showed no spoilage after 28 days of storage and maintained their acceptability.¹⁴ This demonstrates that the combined use of liquid smoke and nanoemulsion, without artificial preservatives, significantly extends the microbiological shelf life of SV mussel meat. Furthermore, since non-proteolytic strains of C. botulinum have the potential to produce toxins within 31 days at 3.3 °C cold storage, the shelf life of SV-processed foods is generally limited to 28 days.²²

These microbial proliferation dynamics exhibited strong biological consistency with the monitored physicochemical, textural, and sensory deterioration patterns. Specifically, the exponential outgrowth of TMAB and TPB populations observed after day 12 in the control and SS groups directly accelerated TVB-N formation (Fig. 2), driven by bacterial deaminases and decarboxylases catabolizing free amino acids and non-protein nitrogen compounds into volatile basic amines. In contrast, the prolonged microbial lag phase sustained in the DNS group effectively restricted this metabolic pathway, keeping TVB-N well within the acceptable 35 mg N/100 g threshold throughout storage. Furthermore, post-mortem bacterial metabolism contributed to slight tissue alkalinization toward late storage, while extracellular bacterial proteases, acting in concert with endogenous autolysis, accelerated the cleavage of myofibrillar networks (as corroborated by FT-IR amide band alterations and SEM fibrillar disruption; Fig. 4 and 5). This structural degradation was macroscopically manifested as progressive softening and significant losses in TPA hardness and chewability (Fig. 3). As will be comprehensively discussed in Section 3.9, these interrelationships and the protective delay achieved by the DNS treatment were further corroborated by the multivariate PCA ordination and trajectory pathways (Fig. 9 and 10).

3.8. Sensory analysis

3.8.1. Raw mussel sensory analysis

Appearance is one of the important factors affecting consumer purchasing decisions. The sensory analysis results for raw mussel meat are presented in Fig. 7. According to the evaluation results, raw mussel meat received high liking scores out of 5. The highest values were obtained for the fresh smell (4.76 ± 0.60), sea/seaweed smell (4.72 ± 0.54) and distinctive smell (4.56 ± 0.56) parameters. This indicates that the mussel meat retained its freshness and that the product's characteristic sea smell was positively evaluated by the panelists.

In terms of visual characteristics, the high scores for brightness (4.30 ± 0.76) and wetness (4.12 ± 1.09) reveal that the product surface has a vivid and fresh appearance. In terms of color, the orange color (3.84 ± 1.03) received a lower score compared to other parameters, suggesting that the panelists perceived slight differences in color tone between the samples. The main reason for the color differences observed in the mussel meat is gender difference. Female mussels are generally orange-reddish, while male mussels are light yellow-beige in color. This may be primarily due to differences in gonad development and the amount of carotenoid pigments. Carotenoids are the main cause of color differences in the flesh tissue of aquatic organisms.⁸⁹ Environmental factors such as dietary differences, seasonal changes, and water temperature also play a role. In terms of textural properties, elasticity, firmness, and surface smoothness (4.04 ± 1.02) were found to be at a medium–high level. These results indicate that the tissue integrity of mussel meat, which has a high water content, is preserved and that it is fresh. Overall, the sensory evaluation scores of the raw mussel meat samples were above 3.8, indicating that the product is of high quality in terms of appearance, smell, and texture.

3.8.2. Sensory analysis of cooked mussel meat in SV treatment groups

Sensory analysis results during storage in the study groups and while the overall sensory evaluation results are presented in Fig. 8.

Smoke aroma and smoke odour evaluations were only performed in the SS and DNS groups. In the salty, sweet, and sour taste evaluation, the assessment was based on the degree of taste, and the scores given to these criteria in the spoilage parameters were not evaluated. When calculating the sensory overall assesment scores for the mussel groups to which SV was applied, the criteria common to all groups and changing during storage were considered: appearance, fresh smell, sea/seaweed smell, umami taste, firmness, chewiness, and juiciness scores. On a 5-point hedonic scale, (5) I liked it very much, (4) I liked it, (3) I liked it moderately, (2) I did not like it (consumability threshold value), (1) I did not like it at all. Products with a score of 2 points or below were considered inedible.

Appearance is one of the most important factors determining consumer preference and purchase of ready-to-eat products. When mussel meat is exposed to high temperature and prolonged heat treatment, shrinkage and reduction in size are observed in the meat texture. Therefore, SV was applied in the study to prevent meat shrinkage. When the groups' scores for freshness, smoke odour, and firmness were examined, the scores of the SS and DNS groups, which contained liquid smoke, did not fall below 3 until the day 21. Liquid smoke positively affected the appearance of mussel meat in both direct and nanoemulsion forms. SV technology may help myoglobin remain in the deoxyhaemoglobin form, which is more resistant to heat, by providing an oxygen-free environment, thus ensuring that the product's color remains more appealing and consistent compared to traditional boiling. These sensory appearance scores closely mirrored the instrumental color coordinates (L* and b*), where the emulsion of liquid smoke in DNS attenuated surface yellowing and preserved luminosity by limiting surface dehydration and lipid auto-oxidation.

At the start of storage (day 0), the overall evaluation scores for all groups were in the range of 4.0–4.5, indicating that the products were fresh and of good quality. As the storage period progressed, a gradual decrease in the overall evaluation scores was observed in all groups. Until day 12, the DNS group showed higher scores compared to the other groups. This suggests that the combined application of smoke flavouring and nanoemulsion contributed positively to the sensory stability of the product. From day 18 onwards, a marked downward trend emerged in all groups, and by day 21, the scores approached the consumability threshold of 2.0–2.5. On the day 21, there was no statistically significant difference between the SS, ANS, and DNS groups (p > 0.05). By day 24, all groups had declined to the 1.0–1.5 range and were considered sensorially unconsumable. Sensory analysis was not performed on day 27. In conclusion, it was determined that the overall level of mussel meat cooked using the SV method decreased during storage; however, the DNS group maintained its sensorial quality for a longer period. Similarly, Amin and Abouzied¹⁶ reported that both hot and cold smoked mussels prepared with wood smoke and liquid smoke received high scores (“very good”) from panelists in terms of aroma and taste profiles.

This temporal sensory degradation exhibited high biochemical and physical consistency with monitored physicochemical, textural, and microbial indicators. When the freshness aroma scores of the groups were examined, the aroma scores of the SS and DNS groups, which contained liquid smoke aroma, were high until the day 21(p < 0.05). Durmuş³² reported that nanoemulsions prepared from citrus (orange, mandarin, grapefruit, lemon) essential oils eliminated the dominant fishy odour in rainbow trout fillets and provided a pleasant odour that was more preferred by panelists. In present study, the sensory rejection observed after day 18 in the control group directly coincided with TVB-N accumulation surpassing the 35 mg N/100 g threshold and psychrotrophic bacterial populations exceeding 6.0 log CFU g⁻¹. At this stage, bacterial deamination and enzymatic proteolysis generated offensive volatile amines, organic sulfides, and ammoniacal notes that masked the natural marine aroma. Liquid smoke also eliminated the dominant metabolite odour released during the spoilage process, and it was emphasised that the groups containing liquid smoke were more pleasant to the panelists. Numerous plant products such as thyme, rosemary, citrus, bay leaf, sage, etc. have been used in nanoemulsions to suppress fish odour. The combination of liquid smoke and nanoemulsion was tested for the first time, and DNS odour scores were found to be higher than ANS during storage.

Furthermore, the panel's sensory perception of textural deterioration (perceived softening, reduced elasticity, and diminished chewiness) was directly supported by the instrumental TPA measurements and microstructural findings. The decline in firmness scores past day 18 aligned with the significant drops in instrumental hardness and chewiness (Fig. 3). At the molecular level, this loss of firmness corresponds to the progressive scission of peptide backbones, as evidenced by the flattening and broadening of the amide I and II peaks in the FT-IR spectra (Fig. 5) and the detachment of muscle fiber bundles observed in SEM micrographs (Fig. 4). The ability of the DNS treatment to preserve structural firmness longer than the control is attributed to the protective action of nanoemulsified smoke phenolics against both bacterial enzymatic proteolysis and post-mortem endogenous muscle autolysis.

3.9. Principal component analysis (PCA)

The purpose of utilising principal component analysis (PCA) was to facilitate the simplification of the dataset and to facilitate the exploration of relationships among quality-related variables. The study concentrated on seven indicators: OA, TVB-N, FFAs, TMAB, TAN, TPB, and TYM. Prior to the implementation of PCA, a standardisation process was employed to ensure a level playing field and equitable contribution from all variables. The initial two components (PC1 and PC2) were the focal point of the study, as they demonstrated a high degree of variation. The PCA biplots facilitated the visualisation of the relationship between quality parameters and sample distribution across processing groups (Fig. 9).

The loadings on the variables clarify the two principal axes. PC1 (76.5% of variance) displayed high, similar positive loadings for the four microbial counts: TMAB (0.42), TPB (0.42), TAN (0.42), and TYM (0.39), along with FFAs (0.34), and a strong negative loading for OA (−0.40). This axis generally indicates microbial and lipolytic spoilage that increases with storage time. PC2 (11.7% of variance) was almost entirely influenced by TVB-N (0.97), with other variables near zero, representing the buildup of volatile nitrogenous bases largely independent of microbial growth. Consequently, samples separated primarily along PC1 based on microbial and free-fatty-acid spoilage, and along PC2 based on TVB-N levels.

PCA was employed to examine the alterations in quality that occurred during the storage period in mussels, with the parameters OA, TVB-N, FFAs, TMAB, TAN, TPB and TYM being of particular significance. As indicated by the PCA results, the first principal component (PC1) accounted for 76.5% of the total variance and exhibited robust positive correlations with TVB-N, TMAB, TPB, TYM, and TAN. This finding suggests that the deterioration in quality is predominantly driven by an increase in microbial and nitrogenous compounds. The second principal component (PC2) explained 11.7% of the total variance and was found to be partially related to OA and, to a lesser extent, FFAs; this axis represents oxidative changes. Despite the partial overlap of the groups in the PCA space, the distribution of samples along PC1 indicates that the rate and direction of deterioration varied among the different processing groups. The substantial loading observed along PC1 in the PCA biplot evidently demonstrates that quality loss in mussels is predominantly associated with microbial activity and the breakdown of protein/nitrogenous compounds. The observation that TVB-N, TMAB, TPB, TYM, and TAN are collinear suggests that these parameters reflect a common deterioration mechanism during storage and are highly intercorrelated. Conversely, the separation of OA along the PC2 dimension indicates that oxidative processes may occur partially independently of microbial deterioration. The absence of complete separation between processing groups in PCA space suggests that the applied processes did not fundamentally alter the degradation mechanism. However, different distribution patterns along PC1 imply that these processes may affect the degradation rate and progression dynamics. The findings indicate that the process of quality loss is governed by a specific biochemical axis, with oxidative processes contributing to a secondary and complementary role.

3.10. PCA trajectory analysis

The present study employed PCA trajectory analysis to monitor the deterioration of quality over time during the storage period. The combination of PCA scores with storage duration and processing groups was utilised to visualise the movement of samples in PCA space over time. The trajectory plots provide a visual representation of the changes in quality over time for each treatment, with PCA scores plotted against the time variable. Utilising the ggplot2 software, we maintained color and group consistency, thereby facilitating the discernment of disparities in the rate and direction of degradation.

PCA trajectory analysis revealed distinct temporal degradation pathways among the processing groups during storage, despite their partial overlap in the PCA space. Along the primary degradation axis (PC1, 76.5%), the control group exhibited the most pronounced and irregular progression, characterised by sharp directional changes and a wider dispersion over time, indicating rapid and unstable quality deterioration (Fig. 10). In contrast, the SS group exhibited a more protracted yet uniform trajectory along PC1, indicating a moderate yet uninterrupted progression of spoilage. The ANS and DNS groups exhibited comparatively shorter and more compact trajectories, remaining closer to the origin and demonstrating limited displacement along PC1 throughout storage. Variations along PC2 (11.7%) were less pronounced but highlighted transient deviations associated with oxidative-related changes, particularly in the control and SS groups. The trajectory patterns indicate that processing treatments influenced not only the extent but also the temporal dynamics of quality degradation.

The PCA trajectory plot provides critical insight into the dynamic nature of quality deterioration that cannot be inferred from static PCA biplots alone. The pronounced and erratic movement of the control group along PC1 reflects accelerated microbial and nitrogenous spoilage processes, consistent with the absence of any stabilising treatment. The SS group, while still progressing along the main degradation axis, exhibited a more coherent trajectory, suggesting partial mitigation of spoilage kinetics rather than a fundamental alteration of the degradation mechanism. It is noteworthy that the ANS and DNS treatments exhibited markedly constrained trajectories, indicating a stabilising effect on overall quality evolution and a slower progression of dominant spoilage processes. The presence of limited oscillations along PC2 further indicates that oxidative changes may have played a secondary and more regulated role under these treatments. Collectively, these findings demonstrate that while all groups share a common underlying degradation structure, processing treatments significantly modulate the rate and temporal pattern of quality loss, with DNS and ANS showing the greatest potential for maintaining quality stability during storage.

3.11. Industrial scalability and economic feasibility perspectives

Although a comprehensive techno-economic cost-benefit assessment was beyond the scope of this laboratory-scale study, the proposed hurdle technology demonstrates strong potential for industrial adoption. The formulation utilizes commercially available, economical, and food-grade ingredients (sunflower oil, Tween 80, and liquid smoke) without relying on costly synthetic preservatives or specialized equipment. The sous-vide cooking and packaging steps are already fully compatible with existing commercial seafood processing lines. Economically, the marginal cost introduced by the nanoemulsion preparation (minor surfactant, carrier oil, and energy for ultrasonic dispersion) is expected to be substantially offset by the tangible commercial benefits: extending the microbiological shelf life of ready-to-eat sous-vide mussels by 6 days (from 18 to 24 days). This extended storage window minimizes post-harvest supply chain losses, broadens chilled distribution networks to distant inland markets, and provides a clean-label, value-added gourmet seafood product that commands higher consumer willingness-to-pay. Future studies focusing on pilot-scale continuous production lines and precise techno-economic feasibility analyses will further facilitate commercial transition.

4. Conclusion

This study demonstrates that the synergistic combination of sous-vide (SV) thermal processing and nanoemulsion-based liquid smoke (DNS) constitutes a highly effective hurdle approach for preserving Mediterranean mussel (Mytilus galloprovincialis) meat during chilled storage. The DNS formulation successfully mitigated microbial proliferation, maintaining total mesophilic aerobic and psychrotrophic counts below critical spoilage levels (<6.0 log CFU g⁻¹) and significantly retarding volatile base nitrogen (TVB-N) accumulation within acceptable regulatory limits (<35 mg N/100 g) for up to 21–24 days, compared to only 12–15 days for the untreated control. Importantly, these biochemical and microbiological preservations were directly mirrored in sensory quality attributes. The nanoemulsion of liquid smoke ensured the controlled release of phenolic fractions, which masked off-odors and off-flavors caused by spoilage metabolites without imparting excessive sensory pungency. Consequently, the DNS group received significantly higher freshness aroma, pleasant smoke odor, and overall acceptability scores, sustaining consumer-grade quality up to day 21. At the structural and molecular levels, FT-IR spectral shifts in the amide I and II regions, together with SEM observations, confirmed that nanoemulsion smoke constituents protected myofibrillar protein backbones against post-mortem autolytic and microbial degradation. This structural resilience was objectively corroborated by the retention of instrumental firmness and chewiness in texture profile analysis (TPA). Overall, multivariate PCA ordination consolidated these multidisciplinary indices, proving that interfacial nanoemulsion delivery of liquid smoke coupled with SV cooking delivers a viable, clean-label preservation strategy capable of significantly extending the shelf life, textural integrity, and organoleptic appeal of ready-to-eat bivalve products.

Author contributions

Demet Kocatepe: project administration, conceptualization, formal analysis, funding acquisition, investigation, methodology, writing – review and editing, writing – original draft, visualization and supervision. Bengünur Çorapcı: conceptualization, methodology, resources, writing – original draft, Bayram Köstekli: conceptualization, formal analysis, methodology, writing – review and editing, visualization. Can Okan Altan: investigation, visualization, resources, writing – review and editing. Sonay Başkan: formal analysis, investigation. M. Yeşim Çelik: formal analysis, investigation, writing – review and editing.

Conflicts of interest

The authors declare no competing interest.

Acknowledgments

This work was supported by the Sinop University Scientific Research Coordination Unit, Project Number: SÜF-1901-24-007. Ethical committee approval was obtained from the Sinop University Human Research Ethics Committee on 05.09.2024/237 for the cooked mussel sensory analyses. Also, we would like to thank to the Samsun and Sinop Provinces Aquaculture Producers' Association [Samsun ve Sinop İlleri Su Ürünleri Yetiştiricileri Üretici Birliği (SSYUB)] and the Kızılırmak Seafood Industry and Trade Company (Samsun/Türkiye) for their logistical support in delivering the fish to the laboratory in cold conditions, ensuring that it remained fresh after being caught.

Data availability

All data generated or analysed during this study are included in this published article.

References

Peycheva K. Panayotova V. Stancheva R. Makedonski L. Merdzhanova A. Cammilleri G. Ferrantelli V. Calabrese V. Cicero N. Fazio F. Food Sci. Nutr. 2022;10(9):3052–3061. doi: 10.1002/fsn3.2903. [DOI] [PMC free article] [PubMed] [Google Scholar]
FAO, World Food and Agriculture – Statistical Yearbook 2025, FAO, Rome, 2025, 10.4060/cd4313en [DOI] [Google Scholar]
TÜİK, Fisheries and Aquaculture Statistics, https://data.tuik.gov.tr/Search/Search?text=bal%C4%B1k&dil=1, accessed September, 2026
Palamae S. Patil U. Saetang J. Detcharoen M. Suyapoh W. Ma L. Zhang B. Benjakul S. Food Control. 2025;173:111206. [Google Scholar]
Xin X. Bissett A. Wang J. Gan A. Dell K. Baroutian S. Food Control. 2021;124:107874. [Google Scholar]
Karayannakidis P. D. Petridis D. Galinou-Mitsoudi S. Tsiaras S. A. Chatziantoniou S. E. Avdelas L. Papaharisis L. J. Aquat. Food Prod. Technol. 2024;33(9):697–713. [Google Scholar]
Odeyemi O. A. Burke C. M. Bolch C. C. J. Stanley R. Food Res. Int. 2019;121:568–576. doi: 10.1016/j.foodres.2018.12.017. [DOI] [PubMed] [Google Scholar]
Bongiorno T. Tulli F. Comi G. Sensidoni A. Andyanto D. Iacumin L. LWT--Food Sci. Technol. 2018;91:117–124. [Google Scholar]
Jeon E. B. Choi M. S. Kim J. Y. Ha K. S. Kwon J. Y. Jeong S. H. Lee H. J. Jung Y. J. Ha J. H. Park S. Y. Food Control. 2020;109:106954. [Google Scholar]
Kathuria D. Dhiman A. K. Attri S. Trends Food Sci. Technol. 2022;119:57–68. [Google Scholar]
Bernárdez M. Pastoriza L. Food Control. 2011;22(2):257–265. [Google Scholar]
English M. M. Keough J. M. McSweeney M. B. Razul M. S. G. J. Food Sci. 2019;84(6):1547–1553. doi: 10.1111/1750-3841.14642. [DOI] [PubMed] [Google Scholar]
Cavalheiro D. Schmidt F. C. Rodrigues L. G. G. Siga C. Leitempergher F. Laurindo J. B. J. Food Process Eng. 2013;36(2):192–201. [Google Scholar]
Russo G. L. Langellotti A. L. Buonocunto G. Puleo S. Di Monaco R. Anastasio A. Vuoso V. Smaldone G. Baselice M. Capuano F. Masi P. Foods. 2023;12(15):2900. doi: 10.3390/foods12152900. [DOI] [PMC free article] [PubMed] [Google Scholar]
Turan H. Sönmez G. Çelik M. Y. Yalcin M. Kaya Y. J. Food Process. Preserv. 2008;32(6):912–922. [Google Scholar]
Amin H. Abouzied A. S. Aquat. Sci. Fish Resour. 2024;5(1):1–11. [Google Scholar]
Şengör G. F. Kalafatoğlu H. Gün H. Turk. J. Vet. Anim. Sci. 2004;28(5):793–797. [Google Scholar]
Köşker A. R. Gündoğdu S. Ayas D. Bakan M. J. Food Compos. Anal. 2022;106:104326. [Google Scholar]
Bhunia K. Ovissipour M. Rasco B. Tang J. Sablani S. S. J. Sci. Food Agric. 2017;97(1):324–332. doi: 10.1002/jsfa.7738. [DOI] [PubMed] [Google Scholar]
Zhan Y. Tu C. Jiang H. Benjakul S. Ni J. Dong K. Zhang B. Foods. 2022;11(23):3928. doi: 10.3390/foods11233928. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bekhit A. E.-D. A. and Roohinejad S., in Reference Module in Food Science, ed. G. W. Smithers, Elsevier, 2016, pp. 1–10 [Google Scholar]
El-Ansari A. and Bekhit A. E.-D. A., in Minimally Processed Foods: Technologies for Safety, Quality, and Convenience, ed. M. W. Siddiqui and M. S. Rahman, Springer, Cham, 2015, pp. 125–150 [Google Scholar]
Mansur D. Sugiwati S. Rizal W. A. Suryani R. Maryana R. Biomass Convers. Biorefin. 2023;13(12):10561–10574. doi: 10.1007/s13399-021-01896-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ali F., Cundari L., Miskah S. and Prasetyo H., in Proceedings of the 4th Forum in Research, Science, and Technology (FIRST-T1-T2-2020), Atlantis Press, 2021, pp. 328–338 [Google Scholar]
Da Porto C. Moret S. Soldera S. Anal. Chim. Acta. 2006;563(1–2):396–400. [Google Scholar]
Donsì F. Ferrari G. J. Biotechnol. 2016;233:106–120. doi: 10.1016/j.jbiotec.2016.07.005. [DOI] [PubMed] [Google Scholar]
Uçar Y. Eur. J. Lipid Sci. Technol. 2020;122(5):1900405. [Google Scholar]
Meral R. Ceylan Z. Köse Ş. J. Food Saf. 2019;39(4):e12644. [Google Scholar]
Corapci B. Turan H. Altan C. O. Kostekli B. Ceylan Z. Kocatepe D. Eur. J. Lipid Sci. Technol. 2025:e70055. doi: 10.1007/s13197-025-06352-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Corapci B. Kocatepe D. Altan C. O. Ceylan Z. Kostekli B. Turan H. J. Food Sci. Technol. 2025;62:1676–1688. doi: 10.1007/s13197-025-06352-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Takadaş F. Özoğlu F. Uçar Y. Kuley E. Durmuş M. Köşker A. R. Sakarya Y. Özkütük S. T. Yılmaz N. Zhang Z. Özgül Y. Food Biosci. 2026;78:108459. [Google Scholar]
Durmuş M. J. Food Saf. 2020;40(1):e12718. [Google Scholar]
Hamouda T. Hayes M. M. Cao Z. Tonda R. Johnson K. Wright D. C. Baker J. R. J. Infect. Dis. 1999;180(6):1939–1949. doi: 10.1086/315124. [DOI] [PubMed] [Google Scholar]
CIE, CIE 015:2018 Colorimetry, International Commission on Illumination (CIE) Central Bureau, Vienna, Austria, 4th edn, 2018 [Google Scholar]
Ludorff W. and Meyer V., Fische und Fischerzeugnisse, Paul Parey, Berlin, 1973, vol. 6 [Google Scholar]
AOAC, Official Methods of Analysis, Association of Official Analytical Chemists, Washington, DC, 16th edn, 1995, pp. 1–75 [Google Scholar]
Oliveira A. R. Sykes A. V. Hachero-Cruzado I. Azeiteiro U. M. Esteves E. Food Chem. 2015;168:520–528. doi: 10.1016/j.foodchem.2014.07.082. [DOI] [PubMed] [Google Scholar]
Çelik M. Y. Karayücel S. Karayücel İ. Öztürk R. Eyüboğlu B. J. Aquat. Food Prod. Technol. 2012;21(3):198–205. [Google Scholar]
Karayücel S. Çelik M. Y. Karayücel İ. Öztürk R. Eyüboğlu B. Aquacult. Res. 2015;46(6):1369–1383. [Google Scholar]
Bibette J., Leal-Calderon F., Schmitt V. and Poulin P., Emulsion Science: Basic Principles. An Overview, Springer, Berlin, Heidelberg, 2002 [Google Scholar]
Sood S. Jain K. Gowthamarajan K. Colloids Surf., B. 2014;113:330–337. doi: 10.1016/j.colsurfb.2013.09.030. [DOI] [PubMed] [Google Scholar]
Corapci B. J. Aquat. Food Prod. Technol. 2022;31(7):672–685. [Google Scholar]
Saupe A. Gordon K. C. Rades T. Int. J. Pharm. 2006;314:56–62. doi: 10.1016/j.ijpharm.2006.01.022. [DOI] [PubMed] [Google Scholar]
Mohammadpour Dounighi N. Eskandari R. Avadi M. R. Zolfagharian H. Mir Mohammad Sadeghi A. Rezayat M. J. J. O. V. A. J. Venomous Anim. Toxins Incl. Trop. Dis. 2012;18:44–52. [Google Scholar]
Özkahraman B. Acar I. Gök M. K. Güçlü G. AKU-J. Sci. Eng. 2014;14:13–21. [Google Scholar]
Kocatepe D. Turan H. Gulgun U. S. F. Corapcı B. Altan C. O. Kostekli B. J. Cul. Sci. Technol. 2025;23(3):649–663. [Google Scholar]
Naveena B. M. Khansole P. S. Kumar M. S. Krishnaiah N. Kulkarni V. V. Deepak S. J. Food Sci. Technol. Int. 2017;23(1):75–85. doi: 10.1177/1082013216658580. [DOI] [PubMed] [Google Scholar]
Palamae S. Temdee W. Buatong J. Zhang B. Hong H. Benjakul S. Innovative Food Sci. Emerging Technol. 2023;87:103391. [Google Scholar]
Dien H. A. Montolalu R. I. Mentang F. Berhimpon S. Nurkolis F. Open Access Maced. J. Med. Sci. 2022;10:1200–1206. [Google Scholar]
Wan J. Cao A. Cai L. Int. J. Gastron. Food Sci. 2019;18:100181. [Google Scholar]
Andlinger D. J. Röscheisen P. Hengst C. Kulozik U. Foods. 2021;10(4):796. doi: 10.3390/foods10040796. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gundermann S., Master's Thesis, The University of Maine, 2023 [Google Scholar]
Chuesiang P. Sanguandeekul R. Siripatrawan U. LWT--Food Sci. Technol. 2020;125:109122. [Google Scholar]
European Commission Commission Decision 95/149/EC. Off. J. Eur. Commun. 1995;L97:84–87. [Google Scholar]
Bimbo A. P., INFORM, 1998, vol. 9, pp. 1–6 [Google Scholar]
Gu M., Master's Thesis, The University of Auckland, 2024 [Google Scholar]
Tóth L. Potthast K. Adv. Food Res. 1984;29:87–158. [Google Scholar]
Martinez O. Salmerón J. Guillén M. D. Casas C. Food Chem. 2007;100(2):498–503. [Google Scholar]
Martinez O. Salmerón J. Guillén M. D. Casas C. Food Control. 2004;15(6):457–461. [Google Scholar]
Pérez-Soto E. Cenobio-Galindo A. d. J. Espino-Manzano S. O. Franco-Fernández M. J. Ludeña-Urquizo F. E. Jiménez-Alvarado R. Zepeda-Velázquez A. P. Campos-Montiel R. G. Molecules. 2021;26(8):2170. doi: 10.3390/molecules26082170. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hultmann L. Rustad T. Food Chem. 2004;87(1):31–41. [Google Scholar]
Hultmann L. Rustad T. Food Chem. 2007;104(4):1687–1697. [Google Scholar]
Prodpran T. Benjakul S. Phatcharat S. Int. J. Biol. Macromol. 2012;51(5):774–782. doi: 10.1016/j.ijbiomac.2012.07.010. [DOI] [PubMed] [Google Scholar]
Hu W. Zhao T. Cao C. Long Y. Zheng J. Huo M. Huo G. Sun W. Su G. Innovative Food Sci. Emerging Technol. 2025;103:104035. [Google Scholar]
Qiu Y. T. Wang Y. M. Yang X. R. Zhao Y. Q. Chi C. F. Wang B. Mar. Drugs. 2019;17(10):565. doi: 10.3390/md17100565. [DOI] [PMC free article] [PubMed] [Google Scholar]
Pongsetkul J. Siriwong S. Thumanu K. Boonanuntanasarn S. Yongsawatdigul J. Foods. 2023;12(3):568. doi: 10.3390/foods12030568. [DOI] [PMC free article] [PubMed] [Google Scholar]
Shahidi F. Dissanayaka C. S. Food Prod., Process. Nutr. 2023;5(1):2. [Google Scholar]
Byler D. M. Susi H. Biopolymers. 1986;25(3):469–487. doi: 10.1002/bip.360250307. [DOI] [PubMed] [Google Scholar]
Farrell Jr H. M. Wickham E. D. Unruh J. J. Qi P. X. Hoagland P. D. Food Hydrocolloids. 2001;15(4–6):341–354. [Google Scholar]
Barth A. Biochim. Biophys. Acta, Bioenerg. 2007;1767(9):1073–1101. doi: 10.1016/j.bbabio.2007.06.004. [DOI] [PubMed] [Google Scholar]
Guillén M. D. Cabo N. J. Am. Oil Chem. Soc. 1997;74(10):1281–1286. [Google Scholar]
Vlachos M. Skopelitis Y. Psaroudaki M. Konstantinidou V. Chatzilazarou A. Tegou E. Anal. Chim. Acta. 2006;573–574:459–465. doi: 10.1016/j.aca.2006.05.034. [DOI] [PubMed] [Google Scholar]
Mezzelani M. Notarstefano V. Panni M. Giorgini E. Gorbi S. Regoli F. Sci. Rep. 2024;14:9369. doi: 10.1038/s41598-024-59663-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kong J. Yu S. Acta Biochim. Biophys. Sin. 2007;39(8):549–559. doi: 10.1111/j.1745-7270.2007.00320.x. [DOI] [PubMed] [Google Scholar]
Kirschner C. Ofstad R. Skarpeid H.-J. Høst V. Kohler A. J. Agric. Food Chem. 2004;52(12):3920–3929. doi: 10.1021/jf0306136. [DOI] [PubMed] [Google Scholar]
Bertram H. C. Kohler A. Böcker U. Ofstad R. Andersen H. J. J. Agric. Food Chem. 2006;54(5):1740–1746. doi: 10.1021/jf0514726. [DOI] [PubMed] [Google Scholar]
Muyonga J. H. Cole C. G. B. Duodu K. G. Food Chem. 2004;86(3):325–332. [Google Scholar]
Carton I. Böcker U. Ofstad R. Sørheim O. Kohler A. J. Agric. Food Chem. 2009;57(9):3563–3570. doi: 10.1021/jf803668e. [DOI] [PubMed] [Google Scholar]
Ozdal T. Capanoglu E. Altay F. Food Res. Int. 2013;51(2):954–970. [Google Scholar]
Zhu Y. Li M. Zhong R. et al. . Food Bioprocess Technol. 2024;17:4739–4753. [Google Scholar]
Lin R. Cheng S. Wang S. Tan M. Zhu B. Foods. 2021;10(2):214. doi: 10.3390/foods10020214. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lytou A. E. Saxton L. Fengou L.-C. Anagnostopoulos D. A. Parlapani F. F. Boziaris I. S. Mohareb F. Nychas G.-J. E. Food Res. Int. 2024;197:115207. doi: 10.1016/j.foodres.2024.115207. [DOI] [PubMed] [Google Scholar]
Wang H. Kay M. Zhang D. Chen G. Li X. Food Chem.:X. 2024;21:101117. doi: 10.1016/j.fochx.2024.101117. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ghazala S., in Improving the Thermal Processing of Foods, ed. P. Richardson, Woodhead Publishing, 2004, pp. 152–174 [Google Scholar]
Tansey F. S. and Gormley T. R., in Novel Food Processing Technologies, ed. G. V. Barbosa-Cánovas, M. S. Tapia and M. P. Cano, CRC Press, 2004, pp. 477–490 [Google Scholar]
Church I. J. Parsons A. L. Int. J. Food Sci. Technol. 1993;28(6):563–574. [Google Scholar]
Chai T. T. Chen Y. W. He F. Y. Jiao R. Cai W. Q. Dong X. P. Liu F. J. Ren S. T. Food Control. 2022;133:108610. [Google Scholar]
Szerman N. Gonzalez C. B. Sancho A. M. Grigioni G. Carduza F. Vaudagna S. R. Meat Sci. 2007;76(3):463–473. doi: 10.1016/j.meatsci.2007.01.001. [DOI] [PubMed] [Google Scholar]
Zheng H. Liu H. Zhang T. Wang S. Sun Z. Liu W. Li Y. Food Chem. 2010;122(4):1164–1167. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All data generated or analysed during this study are included in this published article.

[cit1] Peycheva K. Panayotova V. Stancheva R. Makedonski L. Merdzhanova A. Cammilleri G. Ferrantelli V. Calabrese V. Cicero N. Fazio F. Food Sci. Nutr. 2022;10(9):3052–3061. doi: 10.1002/fsn3.2903. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit2] FAO, World Food and Agriculture – Statistical Yearbook 2025, FAO, Rome, 2025, 10.4060/cd4313en [DOI] [Google Scholar]

[cit3] TÜİK, Fisheries and Aquaculture Statistics, https://data.tuik.gov.tr/Search/Search?text=bal%C4%B1k&dil=1, accessed September, 2026

[cit4] Palamae S. Patil U. Saetang J. Detcharoen M. Suyapoh W. Ma L. Zhang B. Benjakul S. Food Control. 2025;173:111206. [Google Scholar]

[cit5] Xin X. Bissett A. Wang J. Gan A. Dell K. Baroutian S. Food Control. 2021;124:107874. [Google Scholar]

[cit6] Karayannakidis P. D. Petridis D. Galinou-Mitsoudi S. Tsiaras S. A. Chatziantoniou S. E. Avdelas L. Papaharisis L. J. Aquat. Food Prod. Technol. 2024;33(9):697–713. [Google Scholar]

[cit7] Odeyemi O. A. Burke C. M. Bolch C. C. J. Stanley R. Food Res. Int. 2019;121:568–576. doi: 10.1016/j.foodres.2018.12.017. [DOI] [PubMed] [Google Scholar]

[cit8] Bongiorno T. Tulli F. Comi G. Sensidoni A. Andyanto D. Iacumin L. LWT--Food Sci. Technol. 2018;91:117–124. [Google Scholar]

[cit9] Jeon E. B. Choi M. S. Kim J. Y. Ha K. S. Kwon J. Y. Jeong S. H. Lee H. J. Jung Y. J. Ha J. H. Park S. Y. Food Control. 2020;109:106954. [Google Scholar]

[cit10] Kathuria D. Dhiman A. K. Attri S. Trends Food Sci. Technol. 2022;119:57–68. [Google Scholar]

[cit11] Bernárdez M. Pastoriza L. Food Control. 2011;22(2):257–265. [Google Scholar]

[cit12] English M. M. Keough J. M. McSweeney M. B. Razul M. S. G. J. Food Sci. 2019;84(6):1547–1553. doi: 10.1111/1750-3841.14642. [DOI] [PubMed] [Google Scholar]

[cit13] Cavalheiro D. Schmidt F. C. Rodrigues L. G. G. Siga C. Leitempergher F. Laurindo J. B. J. Food Process Eng. 2013;36(2):192–201. [Google Scholar]

[cit14] Russo G. L. Langellotti A. L. Buonocunto G. Puleo S. Di Monaco R. Anastasio A. Vuoso V. Smaldone G. Baselice M. Capuano F. Masi P. Foods. 2023;12(15):2900. doi: 10.3390/foods12152900. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit15] Turan H. Sönmez G. Çelik M. Y. Yalcin M. Kaya Y. J. Food Process. Preserv. 2008;32(6):912–922. [Google Scholar]

[cit16] Amin H. Abouzied A. S. Aquat. Sci. Fish Resour. 2024;5(1):1–11. [Google Scholar]

[cit17] Şengör G. F. Kalafatoğlu H. Gün H. Turk. J. Vet. Anim. Sci. 2004;28(5):793–797. [Google Scholar]

[cit18] Köşker A. R. Gündoğdu S. Ayas D. Bakan M. J. Food Compos. Anal. 2022;106:104326. [Google Scholar]

[cit19] Bhunia K. Ovissipour M. Rasco B. Tang J. Sablani S. S. J. Sci. Food Agric. 2017;97(1):324–332. doi: 10.1002/jsfa.7738. [DOI] [PubMed] [Google Scholar]

[cit20] Zhan Y. Tu C. Jiang H. Benjakul S. Ni J. Dong K. Zhang B. Foods. 2022;11(23):3928. doi: 10.3390/foods11233928. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit21] Bekhit A. E.-D. A. and Roohinejad S., in Reference Module in Food Science, ed. G. W. Smithers, Elsevier, 2016, pp. 1–10 [Google Scholar]

[cit22] El-Ansari A. and Bekhit A. E.-D. A., in Minimally Processed Foods: Technologies for Safety, Quality, and Convenience, ed. M. W. Siddiqui and M. S. Rahman, Springer, Cham, 2015, pp. 125–150 [Google Scholar]

[cit23] Mansur D. Sugiwati S. Rizal W. A. Suryani R. Maryana R. Biomass Convers. Biorefin. 2023;13(12):10561–10574. doi: 10.1007/s13399-021-01896-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit24] Ali F., Cundari L., Miskah S. and Prasetyo H., in Proceedings of the 4th Forum in Research, Science, and Technology (FIRST-T1-T2-2020), Atlantis Press, 2021, pp. 328–338 [Google Scholar]

[cit25] Da Porto C. Moret S. Soldera S. Anal. Chim. Acta. 2006;563(1–2):396–400. [Google Scholar]

[cit26] Donsì F. Ferrari G. J. Biotechnol. 2016;233:106–120. doi: 10.1016/j.jbiotec.2016.07.005. [DOI] [PubMed] [Google Scholar]

[cit27] Uçar Y. Eur. J. Lipid Sci. Technol. 2020;122(5):1900405. [Google Scholar]

[cit28] Meral R. Ceylan Z. Köse Ş. J. Food Saf. 2019;39(4):e12644. [Google Scholar]

[cit29] Corapci B. Turan H. Altan C. O. Kostekli B. Ceylan Z. Kocatepe D. Eur. J. Lipid Sci. Technol. 2025:e70055. doi: 10.1007/s13197-025-06352-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit30] Corapci B. Kocatepe D. Altan C. O. Ceylan Z. Kostekli B. Turan H. J. Food Sci. Technol. 2025;62:1676–1688. doi: 10.1007/s13197-025-06352-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit31] Takadaş F. Özoğlu F. Uçar Y. Kuley E. Durmuş M. Köşker A. R. Sakarya Y. Özkütük S. T. Yılmaz N. Zhang Z. Özgül Y. Food Biosci. 2026;78:108459. [Google Scholar]

[cit32] Durmuş M. J. Food Saf. 2020;40(1):e12718. [Google Scholar]

[cit33] Hamouda T. Hayes M. M. Cao Z. Tonda R. Johnson K. Wright D. C. Baker J. R. J. Infect. Dis. 1999;180(6):1939–1949. doi: 10.1086/315124. [DOI] [PubMed] [Google Scholar]

[cit34] CIE, CIE 015:2018 Colorimetry, International Commission on Illumination (CIE) Central Bureau, Vienna, Austria, 4th edn, 2018 [Google Scholar]

[cit35] Ludorff W. and Meyer V., Fische und Fischerzeugnisse, Paul Parey, Berlin, 1973, vol. 6 [Google Scholar]

[cit36] AOAC, Official Methods of Analysis, Association of Official Analytical Chemists, Washington, DC, 16th edn, 1995, pp. 1–75 [Google Scholar]

[cit37] Oliveira A. R. Sykes A. V. Hachero-Cruzado I. Azeiteiro U. M. Esteves E. Food Chem. 2015;168:520–528. doi: 10.1016/j.foodchem.2014.07.082. [DOI] [PubMed] [Google Scholar]

[cit38] Çelik M. Y. Karayücel S. Karayücel İ. Öztürk R. Eyüboğlu B. J. Aquat. Food Prod. Technol. 2012;21(3):198–205. [Google Scholar]

[cit39] Karayücel S. Çelik M. Y. Karayücel İ. Öztürk R. Eyüboğlu B. Aquacult. Res. 2015;46(6):1369–1383. [Google Scholar]

[cit40] Bibette J., Leal-Calderon F., Schmitt V. and Poulin P., Emulsion Science: Basic Principles. An Overview, Springer, Berlin, Heidelberg, 2002 [Google Scholar]

[cit41] Sood S. Jain K. Gowthamarajan K. Colloids Surf., B. 2014;113:330–337. doi: 10.1016/j.colsurfb.2013.09.030. [DOI] [PubMed] [Google Scholar]

[cit42] Corapci B. J. Aquat. Food Prod. Technol. 2022;31(7):672–685. [Google Scholar]

[cit43] Saupe A. Gordon K. C. Rades T. Int. J. Pharm. 2006;314:56–62. doi: 10.1016/j.ijpharm.2006.01.022. [DOI] [PubMed] [Google Scholar]

[cit44] Mohammadpour Dounighi N. Eskandari R. Avadi M. R. Zolfagharian H. Mir Mohammad Sadeghi A. Rezayat M. J. J. O. V. A. J. Venomous Anim. Toxins Incl. Trop. Dis. 2012;18:44–52. [Google Scholar]

[cit45] Özkahraman B. Acar I. Gök M. K. Güçlü G. AKU-J. Sci. Eng. 2014;14:13–21. [Google Scholar]

[cit46] Kocatepe D. Turan H. Gulgun U. S. F. Corapcı B. Altan C. O. Kostekli B. J. Cul. Sci. Technol. 2025;23(3):649–663. [Google Scholar]

[cit47] Naveena B. M. Khansole P. S. Kumar M. S. Krishnaiah N. Kulkarni V. V. Deepak S. J. Food Sci. Technol. Int. 2017;23(1):75–85. doi: 10.1177/1082013216658580. [DOI] [PubMed] [Google Scholar]

[cit48] Palamae S. Temdee W. Buatong J. Zhang B. Hong H. Benjakul S. Innovative Food Sci. Emerging Technol. 2023;87:103391. [Google Scholar]

[cit49] Dien H. A. Montolalu R. I. Mentang F. Berhimpon S. Nurkolis F. Open Access Maced. J. Med. Sci. 2022;10:1200–1206. [Google Scholar]

[cit50] Wan J. Cao A. Cai L. Int. J. Gastron. Food Sci. 2019;18:100181. [Google Scholar]

[cit51] Andlinger D. J. Röscheisen P. Hengst C. Kulozik U. Foods. 2021;10(4):796. doi: 10.3390/foods10040796. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit52] Gundermann S., Master's Thesis, The University of Maine, 2023 [Google Scholar]

[cit53] Chuesiang P. Sanguandeekul R. Siripatrawan U. LWT--Food Sci. Technol. 2020;125:109122. [Google Scholar]

[cit54] European Commission Commission Decision 95/149/EC. Off. J. Eur. Commun. 1995;L97:84–87. [Google Scholar]

[cit55] Bimbo A. P., INFORM, 1998, vol. 9, pp. 1–6 [Google Scholar]

[cit56] Gu M., Master's Thesis, The University of Auckland, 2024 [Google Scholar]

[cit57] Tóth L. Potthast K. Adv. Food Res. 1984;29:87–158. [Google Scholar]

[cit58] Martinez O. Salmerón J. Guillén M. D. Casas C. Food Chem. 2007;100(2):498–503. [Google Scholar]

[cit59] Martinez O. Salmerón J. Guillén M. D. Casas C. Food Control. 2004;15(6):457–461. [Google Scholar]

[cit60] Pérez-Soto E. Cenobio-Galindo A. d. J. Espino-Manzano S. O. Franco-Fernández M. J. Ludeña-Urquizo F. E. Jiménez-Alvarado R. Zepeda-Velázquez A. P. Campos-Montiel R. G. Molecules. 2021;26(8):2170. doi: 10.3390/molecules26082170. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit61] Hultmann L. Rustad T. Food Chem. 2004;87(1):31–41. [Google Scholar]

[cit62] Hultmann L. Rustad T. Food Chem. 2007;104(4):1687–1697. [Google Scholar]

[cit63] Prodpran T. Benjakul S. Phatcharat S. Int. J. Biol. Macromol. 2012;51(5):774–782. doi: 10.1016/j.ijbiomac.2012.07.010. [DOI] [PubMed] [Google Scholar]

[cit64] Hu W. Zhao T. Cao C. Long Y. Zheng J. Huo M. Huo G. Sun W. Su G. Innovative Food Sci. Emerging Technol. 2025;103:104035. [Google Scholar]

[cit65] Qiu Y. T. Wang Y. M. Yang X. R. Zhao Y. Q. Chi C. F. Wang B. Mar. Drugs. 2019;17(10):565. doi: 10.3390/md17100565. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit66] Pongsetkul J. Siriwong S. Thumanu K. Boonanuntanasarn S. Yongsawatdigul J. Foods. 2023;12(3):568. doi: 10.3390/foods12030568. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit67] Shahidi F. Dissanayaka C. S. Food Prod., Process. Nutr. 2023;5(1):2. [Google Scholar]

[cit68] Byler D. M. Susi H. Biopolymers. 1986;25(3):469–487. doi: 10.1002/bip.360250307. [DOI] [PubMed] [Google Scholar]

[cit69] Farrell Jr H. M. Wickham E. D. Unruh J. J. Qi P. X. Hoagland P. D. Food Hydrocolloids. 2001;15(4–6):341–354. [Google Scholar]

[cit70] Barth A. Biochim. Biophys. Acta, Bioenerg. 2007;1767(9):1073–1101. doi: 10.1016/j.bbabio.2007.06.004. [DOI] [PubMed] [Google Scholar]

[cit71] Guillén M. D. Cabo N. J. Am. Oil Chem. Soc. 1997;74(10):1281–1286. [Google Scholar]

[cit72] Vlachos M. Skopelitis Y. Psaroudaki M. Konstantinidou V. Chatzilazarou A. Tegou E. Anal. Chim. Acta. 2006;573–574:459–465. doi: 10.1016/j.aca.2006.05.034. [DOI] [PubMed] [Google Scholar]

[cit73] Mezzelani M. Notarstefano V. Panni M. Giorgini E. Gorbi S. Regoli F. Sci. Rep. 2024;14:9369. doi: 10.1038/s41598-024-59663-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit74] Kong J. Yu S. Acta Biochim. Biophys. Sin. 2007;39(8):549–559. doi: 10.1111/j.1745-7270.2007.00320.x. [DOI] [PubMed] [Google Scholar]

[cit75] Kirschner C. Ofstad R. Skarpeid H.-J. Høst V. Kohler A. J. Agric. Food Chem. 2004;52(12):3920–3929. doi: 10.1021/jf0306136. [DOI] [PubMed] [Google Scholar]

[cit76] Bertram H. C. Kohler A. Böcker U. Ofstad R. Andersen H. J. J. Agric. Food Chem. 2006;54(5):1740–1746. doi: 10.1021/jf0514726. [DOI] [PubMed] [Google Scholar]

[cit77] Muyonga J. H. Cole C. G. B. Duodu K. G. Food Chem. 2004;86(3):325–332. [Google Scholar]

[cit78] Carton I. Böcker U. Ofstad R. Sørheim O. Kohler A. J. Agric. Food Chem. 2009;57(9):3563–3570. doi: 10.1021/jf803668e. [DOI] [PubMed] [Google Scholar]

[cit79] Ozdal T. Capanoglu E. Altay F. Food Res. Int. 2013;51(2):954–970. [Google Scholar]

[cit80] Zhu Y. Li M. Zhong R. et al. . Food Bioprocess Technol. 2024;17:4739–4753. [Google Scholar]

[cit81] Lin R. Cheng S. Wang S. Tan M. Zhu B. Foods. 2021;10(2):214. doi: 10.3390/foods10020214. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit82] Lytou A. E. Saxton L. Fengou L.-C. Anagnostopoulos D. A. Parlapani F. F. Boziaris I. S. Mohareb F. Nychas G.-J. E. Food Res. Int. 2024;197:115207. doi: 10.1016/j.foodres.2024.115207. [DOI] [PubMed] [Google Scholar]

[cit83] Wang H. Kay M. Zhang D. Chen G. Li X. Food Chem.:X. 2024;21:101117. doi: 10.1016/j.fochx.2024.101117. [DOI] [PMC free article] [PubMed] [Google Scholar]

[cit84] Ghazala S., in Improving the Thermal Processing of Foods, ed. P. Richardson, Woodhead Publishing, 2004, pp. 152–174 [Google Scholar]

[cit85] Tansey F. S. and Gormley T. R., in Novel Food Processing Technologies, ed. G. V. Barbosa-Cánovas, M. S. Tapia and M. P. Cano, CRC Press, 2004, pp. 477–490 [Google Scholar]

[cit86] Church I. J. Parsons A. L. Int. J. Food Sci. Technol. 1993;28(6):563–574. [Google Scholar]

[cit87] Chai T. T. Chen Y. W. He F. Y. Jiao R. Cai W. Q. Dong X. P. Liu F. J. Ren S. T. Food Control. 2022;133:108610. [Google Scholar]

[cit88] Szerman N. Gonzalez C. B. Sancho A. M. Grigioni G. Carduza F. Vaudagna S. R. Meat Sci. 2007;76(3):463–473. doi: 10.1016/j.meatsci.2007.01.001. [DOI] [PubMed] [Google Scholar]

[cit89] Zheng H. Liu H. Zhang T. Wang S. Sun Z. Liu W. Li Y. Food Chem. 2010;122(4):1164–1167. [Google Scholar]

PERMALINK

Comprehensive control of quality deterioration and shelf-life extension in sous-vide mussels using a nanoemulsion–liquid smoke hurdle approach

Demet Kocatepe

Bengunur Corapci

Bayram Köstekli

Can Okan Altan

Sonay Başkan

Meryem Yeşim Çelik

Abstract

1. Introduction

2. Materials and methods

2.1. Reagents and chemicals

2.2. Mussel material and sample preparation

2.3. Preparation of nanoemulsion

2.4. Formation of experimental groups

Table 1. Experimental groups.

2.5. Analyses

2.5.1. Zeta potential (ZPv), mean zeta size (ZSv), and polydispersity index (PDI)

2.5.2. pH measurement

2.5.3. Colorimetric analyses

2.5.4. Measurement of total volatile basic nitrogen (TVB-N)

2.5.5. Free fatty acids (FFAs) analysis

2.5.6. Texture analysis

2.5.7. SEM (scanning electron microscopy) and FT-IR (Fourier transform infrared spectroscopy) analyses

2.5.8. Microbiological analyses

2.5.9. Sensory analysis

2.5.10. Statistical analysis

3. Results and discussion

3.1. Biometric characteristics of mussels

3.2. Mean zeta size (ZSv), zeta potential (ZPv), and polydispersity index (PDI)

3.3. Color measurement

3.4. pH, TVB-N and FFA results

3.5. Texture analysis results

3.6. SEM and FT-IR analyses results

3.7. Microbiological analysis

3.8. Sensory analysis

3.8.1. Raw mussel sensory analysis

Fig. 7. Sensory analysis results for raw mussel meat.

3.8.2. Sensory analysis of cooked mussel meat in SV treatment groups

3.9. Principal component analysis (PCA)

3.10. PCA trajectory analysis

3.11. Industrial scalability and economic feasibility perspectives

4. Conclusion

Author contributions

Conflicts of interest

Acknowledgments

Data availability

References

Associated Data

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases