Cold plasma treatment as a way to sanitize bovine milk while maintaining full nutritional value, stability, toxicological and microbiological safety

Abstract

The scope of this work is to compare cow’s milk hygienization methods (pasteurization, sterilization, and cold plasma treatment) in terms of milk’s toxicological safety (using Danio rerio larvae), microbiological safety (counts of mesophilic aerobic bacteria, yeasts, and molds), nutritional properties (polyphenols, vitamins: B1, B2, B3, B5, B6, B9, and B12, minerals: Ca, Zn, Se), physicochemical properties (pH, conductivity, particle size, color), colloidal stability (turbidimetric method), and chemical changes in milk samples at the molecular level (FT-IR). Pasteurization was performed at 65 °C for 30 min, sterilization at 110 °C for 30 min in an autoclave, while cold plasma treatment was carried out in a gliding arc reactor using nitrogen or air as the working gas for 5, 10, or 20 min, ensuring that the milk temperature was kept below 36 °C. The studies showed that milk preserved with plasma is toxicologically safe and microbiologically safer than fresh milk. Plasma-hygienized milk also retains a higher concentration of vitamins and minerals than temperature-hygienized milk samples and has physicochemical properties similar to those of fresh milk. The use of a gliding arc reactor for milk treatment is important for potential industrial implementation.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-40450-5.

Introduction

Milk is a widely consumed beverage, providing nourishment, hydration, and immune support¹. Interestingly, during their childhood and adulthood, people drink milk from various species, including cows (81%), buffaloes (15%), goats (2%), sheep (1%), and camels (0.4%)^2,3. Milk is also crucial in producing dairy products like cheese, butter, cream, yogurts, and ice cream⁴ because it contains not only water, fats, proteins, carbohydrates, but also bioactive substances such as vitamins and minerals⁵. While milk benefits health, it can also carry harmful contaminants⁶.

Milk undergoes processing, including pasteurization and ultra-high temperature (UHT) sterilization, to guarantee safety and prolong its shelf life⁷. Pasteurization heats milk to 65–85 °C for 25–30 min to inactivate most bacteria, yeasts, and molds, but not spores⁸. It has a minimal impact on protein quality and mineral content, though it may reduce certain B vitamins⁹ and cause minor flavor changes¹⁰. UHT sterilization rapidly heats milk to 135–150 °C for 2–10 s, effectively killing bacteria but causing partial denaturation of whey proteins and some vitamin loss¹¹. Although traditional methods are cost-effective, there is a necessity to use new non-thermal milk processing techniques that retain the maximum nutritional value, minimize color changes, and ensure a high degree of safety. These techniques include high hydrostatic pressure, pulsed electric fields, ultrasound treatment, and cold atmospheric plasma¹². Plasma is an ionized gas containing different reactive species: ions, free electrons, free radicals, reactive atoms and molecules, as well as radiation¹³. The characteristics of plasma depend on the treatment parameters and the type of the working gas¹⁴. Reactive oxygen species (ROS: e.g., hydroxyl radicals (OH.), ozone (O₃), hydrogen peroxide (H₂O₂), singlet oxygen (¹O₂), superoxide anion (O₂⁻))^15–18 and reactive nitrogen species (RNS: e.g., nitric oxide (NO·), dinitrogen tetroxide (N₂O₄), dinitrogen pentoxide (N₂O₅))^18,19can be found in plasma. These components are responsible for its antimicrobial properties, such as inactivation of bacteria, fungi, and yeasts. Cold plasma is currently being applied in the treatment of various types of liquid substances. In addition to water²⁰, plasma is also used, among other applications, to extend the shelf life of fruit juices²¹. Moreover, the diversity of reactive species in plasma gives this sterilization technique a very high efficiency. This is largely due to the challenges faced by pathogens in developing resistance to all different stress factors induced by plasma^10,22. The primary processes of plasma activity resulting in cell apoptosis include (1) erosion of cell membranes, (2) volatilization of compounds under the influence of plasma, and (3) light-induced release of UV photons combined with degradation of the genetic code^23,24. Nowadays, research on cold plasma and its effects on milk has demonstrated that plasma-treated milk is microbiologically safe^25–27. Moreover, both physical and chemical properties of milk-based products are largely unchanged during plasma processing²⁸. Although literature reports some color variations, many investigations have indicated that these differences are not perceptible to human eyes²⁹. Plasma may contribute to the inactivation of endogenous enzymes such as alkaline phosphatase (a pasteurization indicator) and lipases, which contribute to rancidity^30,31. Cold plasma-treated milk may exhibit increased acidity resulting from the multi-step reactions of ROS and RNS with the milk components²⁵. Changes in the acidity of cold plasma treated liquids can be attributed to several factors, such as: liquid volume, buffering capacity, and the type of working gas¹². Interestingly, both fresh and cold plasma treated milk did not show a significant increase in peroxide value observed in pasteurized milk, which can be regarded as a lipid degradation indicator³². A literature review on plasma-treated milk and dairy products is presented in Table 1S (Supplementary material).

The objective of this study was to compare the methods of cow’s milk hygienization (pasteurization, sterilization, and cold plasma treatment) in terms of its toxicological safety, microbiological and nutritional aspects, colloidal stability, and basic physicochemical parameters. Particular emphasis was placed on optimizing the conditions of milk treatment with cold plasma as an eco-friendly alternative that preserves the complete nutritional quality of milk while enhancing the safety of the consumed product. The key innovation of this study is utilizing zebrafish (Danio rerio) larvae as a toxicological model to assess product safety, which is omitted in other studies concerning the impact of cold plasma on milk. Danio rerio is extensively utilized in toxicological research due to several advantageous characteristics. Their small size and ease of maintenance make them a cost-effective model organism. Additionally, zebrafish produce a large number of offspring, facilitating high-throughput studies. Their embryos are transparent and develop rapidly, allowing real-time observation of developmental processes and toxicological effects. These attributes collectively make zebrafish valuable models for assessing the toxicity of various compounds³³. An important advantage of this study is the presentation of changes in the concentration of vitamins and minerals in milk samples subjected to different methods of hygienization. This study also demonstrates the first application of a gliding arc discharge (GAD) reactor for milk treatment. The reactor is characterized by its simple design and indirect processing, in which the active plasma particles come into contact with the material to be processed as a result of a forced flow of working gas. This allows the process temperature to be lowered and the material to be processed in both stationary and flow systems. Unlike other plasma sources, GAD generates stable plasma at atmospheric pressure, making it more compatible with continuous-flow systems. The elimination of complicated vacuum systems and the possibility of using atmospheric air as the working gas offer promising prospects for the practical application in the dairy industry.

Materials and methods

Milk samples and their preservation

Raw milk from Holstein-Friesian cows, supplied directly by a farmer breeding cattle in the Lublin Voivodeship in Poland, was used in the present study. After the morning milking, a portion of the milk was left untreated (control sample, CM), and the rest of the milk was divided into 100 ml samples. The first sample was subjected to heat (pasteurized) in a multi-station water bath with temperature control to 65 °C, and then the temperature was maintained for half an hour. After that, the milk sample was transferred to an ice bath (sample PM). The second milk sample was subjected to conventional sterilization in an autoclave (HIRAYAMA HV-25, Tokyo, Japan) at 110 °C for 30 min (sample SM). The other milk samples were treated with cold nitrogen plasma for 5, 10, or 20 min using a sliding arc plasma reactor (Fig. 1S, Supplementary material) (samples N5, N10, and N20) and with cold air plasma for the same time intervals as above (samples A5, A10, and A20). The cold plasma treatment times were selected in such a way that the shortest one (5 min) caused a noticeable improvement in the microbiological properties of the sample, and the longest time (20 min) did not cause an excessive increase in the sample temperature (maximum acceptable sample temperature: 36 °C). This selection was based on preliminary studies, in which samples were tested after cold plasma treatment for 30 s, 1 min, 2 min, 5 min, 10 min, 15 min, 20 min, and 30 min.

During plasma treatment, a Gliding Arc Discharge (GAD) reactor operating at atmospheric pressure was used. Copper electrodes, 80 mm in length, 2 mm in diameter, and with a divergence angle of 12°, along which the discharge propagates, were placed inside a glass tube with a diameter of 50 mm to prevent the working gas from escaping the plasma stream area. The reactor was powered by a high-voltage converter operating at a frequency of 20 kHz, with an active power of 40 W and a maximum voltage of 18 kV. The frequency of discharges between the electrodes was 50 Hz. The electrical parameters of the reactor were determined using a Rohde & Schwarz RTB2004 oscilloscope, a Tektronix P6015A voltage probe, and a GW Instek GPM-8310 power meter. During the experiments, the gas flow was controlled using glass panel rotameters (Zakłady Automatyki “ROTAMETR”). The flow rate of the working gases, i.e. air and nitrogen, was 440 l/h. The vessel containing the tested sample was placed on the device’s worktable/base. Inside the tested liquid were stir bars (magnetic dipoles), which were set in motion by a magnetic field generated by a magnetic stirrer. Before being placed in the tested substance, the stir bars were cleaned and sterilized. This setup allows contactless mixing of the substance directly during plasma treatment. In addition, the device enables control of the rotational speed, which was adjusted so that the rotational motion of the substance was clearly observable (250 rpm).

All additional details regarding the reactor characterization and plasma process operation are described in the Supplementary material.

All milk samples were stored in sterile, hermetically sealed containers at 4 °C immediately after processing. This protocol was implemented to ensure microbiological safety and minimize compositional changes prior to analysis. The storage period did not exceed 24 h for any sample. The same handling conditions were applied uniformly across all experimental groups to maintain consistency and ensure the reliability of the results.

Toxicological analysis

All experiments were conducted following the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and the European Community Council Directive for the Care and Use of Laboratory Animals of 22 September 2010(2010/63/EU), which applies in Poland. For the experiment with larvae up to 5-dpf (days post fertilization), approval of the Local Ethical Commission is not required. After each experiment, larvae were immediately anesthetized and sacrificed with 15% tricaine solution. All methods were reported in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines (https://arriveguidelines.org), ensuring transparent and comprehensive reporting of experimental design, animal husbandry, experimental procedures, sample size, statistical methods, and ethical considerations.

Institutional review board statement

Ethical review and approval were waived for this study, because according to the EU Directive 2010/63/EU on the protection of animals used for scientific purposes, the Directive applies only to live non-human vertebrate animals from the stage at which they are capable of independent feeding. In the case of zebrafish (Danio rerio), this stage has been widely recognized to occur at 5 days post-fertilization (dpf). Scientific publications support this developmental benchmark. Additionally, the SCHEER Opinion prepared as scientific guidance for the Directive explicitly states that zebrafish larvae reach the relevant level of maturity at 5 dpf. Therefore, experiments conducted on zebrafish embryos prior to this stage do not fall within the scope of the Directive and are exempt from ethical review.

Animals

Danio rerio of the AB strain were housed at the Experimental Medicine Center, Medical University of Lublin, Poland, in standard aquaculture conditions at 28 °C with a 14-h light and 10-h dark cycle. Fertilized eggs were obtained through natural spawning. Embryos were cultured in E3 embryo medium, with a pH of 7.1–7.3, containing 4.7 mM sodium chloride, 0.17 mM potassium chloride, 0.3 mM calcium chloride, and 0.38 mM magnesium chloride, and incubated at 28.5 °C. The study involved zebrafish larvae at four days post-fertilization (4 dpf). Immediately following the experiment, the larvae were humanely euthanized using a tricaine overdose. All procedures used have been previously documented³⁴.

Mortality and heartbeat of Danio rerio larvae

To evaluate the effect of the milk, the mortality and heart rate of Danio rerio larvae were determined in different dilution ratios (1:150, 1:100, 1:75, 1:50, 1:25) using freshly diluted milk. All details concerning these studies are described in the Supplementary material.

Microbiological studies

The microbial integrity of the studied milk samples was evaluated using surface inoculations performed with the plate method on agar-based media, following standards^35,36. The inoculations were prepared in a laminar flow hood equipped with a UV lamp (CRUMA 670FL, Spain). Aerobic bacteria were cultured on PCA medium at 30 °C for 72 h, while yeasts and molds were grown on YPD medium with dichloran and 18% glycerol (DG18) at 25 °C for 5 days. Samples were incubated in a POL-EKO CLN 115 SMART incubator (Poland). The resulting colonies were quantified, and the findings were reported as log10 colony-forming units per milliliter of milk (log₁₀ CFU/ml).

Physicochemical properties

Vitamin content

The content of vitamins (B1, B2, B3, B5, B6, B9, and B12) in the tested milk samples was determined chromatographically using an LC/MS system consisting of a UHPLC chromatograph (UltiMate 3000, Dionex, Sunnyvale, CA, USA) and a mass spectrometer (LTQ-Orbitrap Velos by Thermo Fisher Scientific, San Jose, CA) equipped with an ESI source. The details concerning these studies are described in the Supplementary material.

Mineral content

Milk mineralization was performed using a CEM Mars Xpress (Matthews, USA), according to a single-phase program: 800 W, temperature 210 °C for 15 min. 4 ml of nitric acid was added to each 0.5 g sample. After completion of the mineralization process, the solutions obtained were diluted with deionized water to a final volume of 50 ml. The concentrations of calcium, zinc, and selenium in the tested milk samples were determined by flame atomic absorption spectrometry (FAAS), using a Varian SpektrAA 280FS spectrometer with an SPS-3 autosampler and a SIPS diluent (Agilent Technologies, California, USA). The mineralization and element determination stages were verified by method validation, taking into account parameters such as linearity (R² > 0.998 in the concentration range of 0.05–5.0 mg/l), limit of detection (LOD) and quantification (LOQ), precision (RSD < 5%), and recovery of analytes in the range of 95–105%. Verification was performed using certified reference materials and analysis of control samples.

Total polyphenol content

Total phenolic content was determined using the Singleton method³⁷with minor modifications as in Barba et al.³⁸. 1.5 ml of distilled water, 2.5 ml of Folin-Ciocalteau reagent, and 2 ml of a 20% Na₂CO₃ solution were added to 300 µl of milk. The mixtures were incubated in the dark for 90 min until a stable characteristic blue color was obtained. Absorbance was then measured spectrophotometrically (Shimadzu UV-1800; Kyoto, Japan) at a wavelength of λ = 725 nm. The calibration curve was developed based on a series of standard solutions prepared from a standard solution of caffeic acid at a concentration of 1.0 mg/ml. To reduce interference with the measurement resulting from natural turbidity and components of the milk matrix absorbing light at 725 nm, each sample was previously centrifuged (e.g. 10 000×g, 10 min.), and then the supernatant obtained was used for the reaction with Folin-Ciocalteau reagent. Additionally, all analyses were performed against a blank sample containing the same proportions of reagents without milk, which allowed correction for background absorbance.

Colloidal stability tests

The stability of the milk samples was evaluated using the Turbiscan LabExpert (L’Union, France) equipped with a TLab Cooler refrigeration module. The sample was poured into a measurement vessel, which was then placed in the device within a temperature-controlled chamber. Each sample was scanned for 72 h at 25 °C (hourly during the first 24 h, then at 48 h and 72 h). All measurements were performed twice. The relationship between the Turbiscan Stability Index (TSI) and time was calculated.

Basic physicochemical analysis

pH was measured using a pH meter (Elmetrin CX-401) by directly dipping the electrode (EPS-1 electrode) into the sample. To evaluate the conductivity of the milk samples, a conductivity meter (Elmetrin CX-401) with the EC-60 electrode was used. Measurements were performed in quadruplicate, and the resulting mean values along with the standard deviation are presented. Color measurements were made with the use of a 3Color spectrophotometer, model SF80 (Marcq-en-Barœul, France) with illuminant D65, 10° observer. To ensure reliable and reproducible color measurements, milk samples were thoroughly mixed prior to analysis to achieve homogeneity, particularly with respect to the distribution of fat and other components. Measurements were performed approximately 15 min after removal from refrigeration, once the samples reached room temperature (~ 20 °C). A clean cuvette with a fixed optical path length of 1 cm, made of non-absorbing optical material, was used for all colour assessments. The space for expressing the color of the milk samples was the CIE LAB with three coordinates: lightness (L*), red/greenness (a*), and yellow/blueness (b*). Calibration was performed using a white and black ceramic plate (L*: 92.37, a*: − 0.82, b*: 1.82).

The total color difference (ΔE) was calculated from the equation:

1

where L_o^*, a_o^*, and b_o^* are the values for the control milk (CM).

Particle size measurements

The particle size distribution in the milk samples was analyzed using two methods: microscopically (optical microscope AZ100M from Nikon, Japan) and using the NanoZS Zetasizer (Malvern Instruments, UK). Microscopic images were taken from at least three different regions of each sample, and one example result is presented. The measurements with the NanoZS Zetasizer were carried out at 25 °C. The refractive index of milk taken into calculations was 1.3463³⁹, whereas the absorption coefficient was 0.1 cm^− 140. All measurements were conducted without sample dilution, in six replicates, and mean values were reported.

FT-IR analysis

FT-IR/ATR spectra were taken in the 4000–1000 cm^− 1 range with a resolution of 4 cm^− 1 on a spectrometer (Jasco FT-IR-4200 Type A FT-IR, Japan) equipped with a single-reflection ATR accessory with a ZnSe crystal (Jasco ATR PRO ONE, Japan). Each spectrum was obtained by averaging 32 scans. Measurements were performed after the milk samples had equilibrated to room temperature, either naturally (control) or after cooling (treated samples). For each sample, the spectra were obtained after an initial background measurement in the shortest possible interval, without additional normalization or post-processing to allow direct quantitative comparison.

Statistical analysis

The studies were conducted in two series, each analysis performed in triplicate. The results were compared and developed using the Statistica 8.0 computer program (StatSoft Inc., Tulsa, OK, USA) and Prism v8.3.1 from GraphPad (Boston, MA, USA), reporting arithmetic means and standard deviations for all measured parameters. In order to characterize the differences between individual milk samples a one-way ANOVA was conducted on quality indicators. In addition, significance between sample means was assessed using Tukey’s test (p ≤ 0.05).

Results and discussion

Toxicological studies

In our study, we used zebrafish (Danio rerio) larvae as a model organism to assess the toxicological safety of milk samples subjected to different processing methods: standard thermal processing and cold plasma treatment in varying conditions. This model is widely used in toxicological research due to several advantages, including physiological and genetic similarity to higher vertebrates, rapid development, embryo transparency, and high sensitivity to toxicants. The rationale for using this model lies in the hypothesis that milk processing may lead to the formation of reactive species or degradation by-products (e.g. oxidized lipids, altered proteins) that may affect biological systems. Thus, by exposing zebrafish larvae to different milk dilutions, we aimed to simulate ingestion-related exposure conditions and evaluate the biological safety and compatibility of the processed milk samples.

To explore the tolerance of zebrafish embryos to different milk preparations, we used five milk dilutions (1:25, 1:50, 1:75, 1:100, and 1:150) and incubated 4 dpf zebrafish larvae in each for 20 h. Changes in the pH of the milk sample dilutions measured at 28 °C immediately after sample preparation and after 24 h are presented in Table 1S. The toxicological studies showed that the exposure did not cause significant mortality for up to 8 h (Table 2S). The observations after 20 h (Fig. 1) revealed that, at the 1:50 dilution, PM caused 100% mortality, whereas CM at the same concentration resulted in only 46.7% mortality. Meanwhile, SM caused 100% mortality at all the tested dilutions. For the cold plasma treated samples in the nitrogen atmosphere, 100% mortality was observed in the 20-min exposure scheme for dilutions of 1:25 and 1:50, while a mortality rate of 93.3% was recorded for the 1:75 dilution. In contrast, for the milk samples treated with cold plasma in the air atmosphere, 100% mortality was observed in the two lowest dilutions (1:25 and 1:50) within the 5-min exposure scheme. Additional data concerning the heart rate of the zebrafish larvae are presented in Fig. 2S. The toxicological studies have indicated that plasma treatment of milk represents a safer alternative to standard processing methods, considering the higher survival rates in zebrafish.

Fig. 1.

Larval viability after 20 h of exposure to different dilutions of the tested formulations. (A) CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; (B) nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; (C) air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

It was observed that nitrogen-generated plasma appeared less toxic than plasma generated in air. This difference likely results from the distinct reactive species produced during treatment: air plasmas tend to produce higher concentrations of reactive oxygen species, which are strongly oxidative and can damage sensitive biomolecules, while nitrogen plasmas predominantly generate reactive nitrogen species, which appear less aggressive in this respect. This reduced oxidative stress could explain the higher survival rates of Danio rerio larvae observed after exposure to nitrogen plasma–treated milk⁴¹. Furthermore, the higher toxicity and lower survival rates in pasteurized and sterilized milk compared to plasma-treated milk can be explained by several heat-induced mechanisms reported in the literature. Thermal processing accelerates Maillard reactions and the formation of advanced glycation end-products, with known cytotoxic effects, promotes lipid oxidation leading to toxic aldehydes, and causes the degradation of heat-sensitive vitamins that normally provide antioxidant protection. In addition, protein denaturation and aggregation during heating may yield poorly digestible or bioactive compounds. In contrast, plasma treatment achieves microbial inactivation without generating such by-products, which likely explains the better toxicological safety observed in the zebrafish model^42,43. These findings support the utility of zebrafish as a sensitive in vivo model for preclinical toxicological screening and provide a statistically robust measure of milk safety beyond physicochemical or microbiological assessments.

Microbiological studies

Table 1 shows the results of our microbiological tests of cow’s milk subjected to different types of processing. The results indicate The results indicate considerable variation in microbial reduction efficiency, depending on the method used, both in terms of total aerobic mesophilic bacterial count (TAMB) and yeasts and molds (YM). In CM, the TAMB and YM levels were 4.23 log CFU/ml and 3.91 log CFU/ml, respectively. Cold plasma treatment caused a measurable reduction in the tested microorganisms compared to the control milk, with bacterial counts decreasing as treatment duration increased. The largest statistically significant difference (p ≤ 0.05) from the CM samples was noticed for the milk sample subjected to cold plasma for 20 min. The TAMB level in samples N20 and A20 was 3.42 log CFU/ml and 3.26 log CFU/ml, respectively. In turn, the YM count in these samples was 2.79 log CFU/ml (N20) and 2.71 log CFU/ml (A20). Milk treated with cold plasma for 5 min, with the nitrogen working gas, showed a TAMB count of 4.18 log CFU/ml and YM of 3.84 log CFU/ml. Using the same treatment duration with air as the working gas caused a slight decrease in these values to 4.13 log CFU/ml (TAMB) and 3.81 log CFU/ml (YM) (Table 1). The highest and almost complete inactivation was observed in the PM and SM samples. Both the TAMB and YM counts were 0 log CFU/ml. This means that the heat treatment (pasteurization and sterilization) effectively eliminated all the microorganisms. Such results are in line with the general principles of pasteurization and sterilization efficiency, which use high temperature to eliminate microorganisms, including more resistant bacteria¹⁰. Comparable results have also been reported in other studies. For example,⁴⁴ observed mesophilic aerobic counts of 0 CFU/ml in pasteurized milk samples, confirming that complete microbial inactivation can be achieved in controlled industrial conditions. Similarly, a study conducted at the Korhogo Dairy reported no presence of yeasts, molds, Salmonella spp., or Staphylococcus aureus in pasteurized milk, supporting the high effectiveness of thermal treatment in microbial elimination⁴⁵. Importantly, the final microbial levels in the plasma-treated samples remained within the microbiological limits defined for milk products. According to the Polish national standard⁴⁶, the acceptable limit for yeasts and molds in milk is typically below 3.00 log CFU/ml. Similarly, the Regulation of the Minister of Agriculture and Rural Development⁴⁷specifies that the total count of mesophilic aerobic bacteria in pasteurized milk should not exceed 4.00 log CFU/ml. Based on these criteria, the N20 and A20 cold plasma-treated samples meet the current microbiological safety and quality standards. Therefore, it can be concluded that cold plasma, particularly with longer exposure times, may serve as an effective non-thermal preservation method that helps milk meet legally required microbiological criteria without compromising its nutritional value.

Table 1.

Total microbial count and yeast & mold count in milk subjected to different processing methods.

Milk samples	Mesophilic aerobic bacterial counts (log CFU/mL)	Yeast and molds (log CFU/mL)
CM	4.23 ± 0.14d	3.91 ± 0.11c
PM	0.00 ± 0.00a	0.00 ± 0.00a
SM	0.00 ± 0.00a	0.00 ± 0.00a
N5	4.18 ± 0.12d	3.84 ± 0.10c
N10	3.47 ± 0.13c	2.83 ± 0.11b
N20	3.42 ± 0.09c	2.79 ± 0.09b
A5	4.13 ± 0.11d	3.81 ± 0.12c
A10	3.35 ± 0.12bc	2.74 ± 0.11b
A20	3.26 ± 0.10b	2.71 ± 0.09b

The results are expressed as a mean ± standard deviation. ^{a, b, c…}. - values in columns with different lowercase superscripts differ significantly (p ≤ 0.05). CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

The results show that the cold plasma treatment, regardless of the gas used, caused a substantial decrease in microbial counts, especially after the longer exposure (10 and 20 min). The higher efficiency in reducing microbial counts may be due to the longer exposure to reactive chemical species that can damage not only the cell membrane but also DNA and other key cellular components¹⁰. In contrast, the increased number of microorganisms in the control milk may indicate the presence of undesirable bacteria and yeasts, which is typical in the absence of appropriate preservation methods⁴⁸. In the case of cold plasma generated with either nitrogen or air as the working gas, a gradual decrease in the number of yeasts and molds is seen, demonstrating the ability of cold plasma to eliminate fungi, which frequently pose problems in milk and milk-derived foods⁴⁹. Moreover, microorganisms exhibit varying susceptibility to cold plasma. Gram-negative bacteria typically exhibit higher sensitivity to plasma treatment compared to gram-positive bacteria, owing to differences in their cell wall structures. This microbial inactivation mechanism complements the preservation of milk’s intrinsic nutritional components, which remain largely unaffected by the non-thermal nature of plasma treatment. Consequently, cold plasma offers the dual benefit of effective microbial reduction while maintaining the milk’s nutritional and sensory qualities, distinguishing it from conventional thermal methods. As indicated by scientific data, non-thermal plasma is efficient in deactivating a broad spectrum of microorganisms, including both pathogenic and spoilage organisms. However, gram-negative bacteria are usually more sensitive to plasma processing than gram-positive ones due to their thinner and less damage-resistant peptidoglycan layer²⁶. Bacterial spores tend to be more resistant to cold plasma treatment, due to their complex multilayer protective structures, which hinder the penetration and inactivation by reactive species generated during the process. In contrast, yeasts and molds show a gradual and significant reduction in numbers following cold plasma exposure, as they lack such robust protection and are susceptible to membrane disruption by plasma-generated reactive compounds. Recent studies and reviews emphasize that the antimicrobial efficiency of cold plasma depends not only on the organism’s taxa but also on the structure and resilience of their cell envelopes, with fungal cells often displaying intermediate sensitivity compared to vegetative bacteria and bacterial spores⁵⁰. Non-thermal plasma treatment offers a distinct advantage over conventional pasteurization techniques, as it operates at ambient temperatures and thus minimizes degradation of milk’s nutritional components⁵¹. The efficacy of this method is influenced by parameters such as exposure duration and the specific gas composition employed during processing⁵².

Cold plasma treatment is attracting increasing interest as a method of microorganism inactivation in a majority of milk-based products. Analyses of the impact of cold plasma on bacteria in milk have shown that plasma significantly reduces the number of pathogens such as E. coli O157:H7. Studies reported a 54% decrease in the bacterial population after just 3 min of plasma exposure, regardless of the fat content²⁹. Non-thermal, cold plasma treatment also resulted in the complete elimination of E. coli in milk after 20 min, which indicates its high effectiveness in inactivating pathogens without significant influence on pH or organoleptic properties of milk²⁹. Other authors²⁵showed that 10-min processing with dielectric barrier discharge (DBD) plasma reduced the number of bacteria and improved the microbiological quality of milk, making this technology appealing in terms of promoting food safety. A benefit of the plasma technology is its capacity to eliminate a wide range of pathogens without using chemicals, which can be of great importance in the context of preserving the quality of food and extending its shelf life. In addition, the possibility of adjusting plasma parameters, such as treatment time, gas type, discharge frequency, and power, allows precise process control of the process to the variety of microorganisms and the specific requirements of food products¹⁰.

Nutritional content results

Based on the toxicological assessment showing the highest zebrafish survival and lowest signs of stress at 10-min plasma exposure, vitamin analysis was performed for N10 and A10 samples, as they represented the optimal plasma treatment conditions balancing safety, microbiological quality, and nutritional preservation. Table 2 presents the vitamin content in the studied milk samples. Milk contains numerous vitamins and minerals beneficial to human health, helping prevent autoimmune, cardiovascular, and even cancerous diseases. In order to understand these health benefits, it is important to consider both the nutritional composition of milk and the microorganisms involved in its digestion. Fat-soluble vitamins, namely A, D, E, and K, as well as water-soluble B-class vitamins, are present in cow’s milk and contribute to its nutritional value. This product is an excellent natural source of vitamin B12 due to the abundance of bacteria capable of producing this nutrient in the cow’s digestive tract⁵³. Vitamins, particularly thiamine (B1), are sensitive to high temperatures. This applies primarily to thiamine (vitamin B1), which is important for the nervous and cardiovascular systems, supports the immune system, and exhibits antioxidant properties. Although effective in extending shelf life, pasteurization and sterilization reduce the concentration of folic acid (B9) through non-enzymatic browning and the degradation of other heat-sensitive vitamins⁵⁴. Table 2 shows that the sterilization of the milk resulted in a decrease in all vitamins included in the research program (compared to the values in the control milk), and the largest losses of over 33% were observed in the case of vitamin B12. Given these limitations of thermal processing, innovative non-thermal methods have been investigated as alternatives to preserve milk quality while ensuring safety and non-thermal processing method may be an alternative to high-temperature milk preservation. In particular, the use of N₂ as a working gas contributed to a reduction of losses in the amount of B vitamins, which did not exceed 11%. However, reactive oxygen species (ROS) generated during cold plasma treatment can interact with vitamins, polyphenols, and milk proteins, potentially causing oxidation and structural modifications. While short exposure times maintain nutritional quality, longer treatments may lead to degradation of sensitive compounds and reduce antioxidant properties. ROS-induced changes in protein structure can also affect water retention and mineral stability in milk. Therefore, treatment duration and plasma parameters are critical for balancing microbial safety and nutritional preservation. These findings are supported by Sharma and Singh (2023)⁵⁵ who showed that cold plasma treatment affects milk protein sulfhydryl content and texture through oxidation reactions, highlighting the importance of controlled treatment time and gas composition for maintaining milk quality.

Table 2.

Vitamin content in the studied milk samples.

Vitamin	CM	PM	SM	A10	N10
	Concentration of vitamin [µg/ml]
B1	0.397 ± 0.020b	0.366 ± 0.018b	0.293 ± 0.013a	0.378 ± 0.018b	0.388 ± 0.018b
B2	1.512 ± 0.082a	1.494 ± 0.075a	1.310 ± 0.061a	1.449 ± 0.070a	1.509 ± 0.070a
B3	1.498 ± 0.080a	1.476 ± 0.081a	1.416 ± 0.066a	1.378 ± 0.064a	1.487 ± 0.067a
B5	3.124 ± 0.182b	2.918 ± 0.146ab	2.506 ± 0.118a	2.990 ± 0.142ab	3.093 ± 0.140b
B6	0.478 ± 0.026b	0.440 ± 0.026b	0.389 ± 0.025a	0.452 ± 0.022b	0.466 ± 0.021b
B9	0.084 ± 0.006b	0.071 ± 0.004ab	0.061 ± 0.004a	0.071 ± 0.004ab	0.075 ± 0.004ab
B12	0.018 ± 0.001b	0.016 ± 0.001b	0.012 ± 0.001a	0.017 ± 0.001b	0.017 ± 0.001b

The results are expressed as a mean ± standard deviation. ^{a, b, c…}. - values in rows with different lowercase superscripts differ significantly (p ≤ 0.05). CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

There is a research gap in the scientific literature regarding the impact of this type of food processing on the content of vitamins in food products. Fragmented data showed an increase in vitamins B3 and B6 in seriguela juice, depending on the processing time and nitrogen gas flow rate in glow discharge plasma. This may have resulted from the release of chemically bound compounds⁵⁶. Reports on vitamin C are ambiguous, suggesting that its increase may result from plasma-produced nitric oxide enhancing dehydroascorbate reductase activity, which reduces oxidized ascorbic acid^57,58 in the presence of free radicals (ROS and RNS)⁵⁹.

The high bioavailability of calcium from dairy products is partly attributed to the presence of high-quality protein in milk, readily digestible with a favorable amino acid profile and an appropriate ratio of essential amino acids⁴⁷. High calcium levels in milk enhance rennet activity, improving curd coagulation, structure, and syneresis^60,61. Milk is also a valuable source of Zn, an important component of enzymes with antioxidant effects. This element also has the ability to counteract the formation of free radicals, which helps reduce the risk of inflammatory, neoplastic, and neurodegenerative diseases. Milk also contains selenium which, although not abundant, is highly bioavailable^62,63.

Table 3 presents the average concentrations of inorganic elements in control milk and remaining samples subjected to different preservation methods. Research indicates that the content of macro- and microelements in milk is variable and depends on many genetic (breed) or environmental (nutrition, season) factors. According to various sources, calcium, zinc, and selenium levels in bovine milk vary widely, ranging from 700 to 1400 mg/kg, 2.3–7.5 mg/kg, and 1.35–1.89 mg/kg, respectively^64,65. In the present study, the control milk contained an average of 1109 mg/kg calcium, 7.60 mg/kg of zinc, and 1.62 mg/kg of selenium. The values of these elements decreased after pasteurization (in the range of 0.7-5%) and sterilization (between 7 and 19%) of the milk. In turn, samples subjected to cold plasma for 10 min, especially with the nitrogen process gas, were characterized by a significantly higher (p ≤ 0.05) content of all the analyzed components in comparison to the CM sample. The other milk samples treated with CAP had similar levels of the macro- and microelements to those in the untreated milk. However, in order to preserve the greatest possible amount of minerals in bovine milk, it is recommended not to extend the CAP treatment with air plasma up to 20 min, as this process can be responsible for a decrease (p ≤ 0.05) in the content of calcium, zinc, and selenium (by approximately 2%, 5%, and 6%, respectively). However, it can be assumed that inorganic compounds are not significantly affected by short CAP treatment times. The presence of calcium in milk has been analyzed by other scientists, who proved that, after treatment with plasma generated in a DBD reactor (continuous non-thermal plasma system) at both air flow rates (3 and 6 ml/min), the content of this macroelement did not differ significantly (p > 0.05) from that in the CM sample and ranged from 1019 to 1100 mg/kg⁶⁶. In general, the results suggest that cold plasma treatment preserves the natural mineral content of milk. The observed preservation or slight increase in mineral concentration after plasma treatment may result from water loss concentration effects. However, ROS and RNS may also interact with milk matrix components, affecting protein conformation and water-binding capacity, which indirectly influences mineral retention. Such effects were reported by Nikmaram and Keener (2022)⁶⁷ who emphasized that plasma-induced oxidative modifications could affect milk’s physicochemical properties while maintaining nutritional quality. The apparent increase in mineral levels may result from a concentration effect caused by water loss from the product. This phenomenon may result from the interaction of plasma-generated ROS and RNS, which alter milk protein structure and matrix components, affecting water retention and physicochemical properties^68,69. On the other hand, Y⁷⁰ found that milk pasteurization at 63 °C for 30 min resulted in almost complete preservation of calcium, while UHT sterilization (130 °C, 2 s) contributed to reducing its amount by almost 5%.

Table 3.

Calcium, zinc, selenium, and total polyphenol content in cow’s milk subjected to various methods of preservation.

Milk samples	Ca (mg/kg)	Zn (mg/kg)	Se (mg/kg)	Polyphenol content (mg/100 ml)
CM	1109.67 ± 0.58de	7.60 ± 0.12cd	1.62 ± 0.03cd	35.37 ± 0.38d
PM	1101.67 ± 2.89c	7.26 ± 0.05b	1.54 ± 0.02b	31.02 ± 0.28c
SM	1003.33 ± 2.89a	7.09 ± 0.08a	1.31 ± 0.01a	15.43 ± 0.45a
N5	1109.67 ± 0.58de	7.51 ± 0.02cd	1.53 ± 0.03b	35.93 ± 0.25de
N10	1139.3 ± 1.15g	7.82 ± 0.03e	1.64 ± 0.04de	36.93 ± 0.25e
N20	1103.33 ± 1.15cd	7.45 ± 0.03c	1.57 ± 0.03bc	29.20 ± 0.25b
A5	1113.33 ± 2.89e	7.62 ± 0.03d	1.64 ± 0.02de	35.13 ± 0.15d
A10	1131.33 ± 2.31f	7.79 ± 0.02e	1.69 ± 0.02e	36.13 ± 0.67de
A20	1085.00 ± 5.00b	7.20 ± 0.02ab	1.51 ± 0.01b	30.03 ± 0.70bc

The results are expressed as a mean ± standard deviation. ^{a, b, c…}. - values in columns with different lowercase superscripts differ significantly (p ≤ 0.05). CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

Phenolic acids, classified as polyphenols, found in many raw materials, such as cereals, carrots, peas, sunflower seeds, nuts, and various fruits, are mostly feed components. It is known that dairy products can be rich in various phytochemicals⁷¹. Lee et al.⁷² observed that, compared to the control group, milk obtained from cows whose standard diet was supplemented with citrus pulp and soybean meal was characterized by significantly higher levels of hesperetin and naringenin as well as 3,4-hydroxybenzoic acid, syringic acid, and caffeic acid. The control milk used in our study had an average total polyphenol content (calculated as caffeic acid) of 35.37 mg/100 ml. Both heat treatments caused a decrease in these compounds by 12–56%. The low-temperature milk preservation resulted in a slight increase in the concentration of the biologically active compounds (sample N10) or a decrease of up to 17% (samples A20 and N20), which was confirmed by the Tukey statistical test (p ≤ 0.05) (Table 3). Calligaris et al.⁷³ observed that HTST pasteurization (72°C, 15 s) caused a decrease in the antioxidant properties of skimmed milk, resulting from the breakdown of natural antioxidants and generation of new oxidizing agents in the Maillard reaction. In turn,⁷⁴ did not report any changes in the polyphenol content (expressed in mg gallic acid equivalents) in skimmed milk powder after 20–120 s CAP treatments using a fluidized bed chamber with nitrogen as the working gas. As found by these researchers, the oxidative stress induced by plasma was not able to change the antioxidant properties of milk. In our case, the 20-min exposure of the milk samples to reactive species occurring during the cold plasma treatment resulted in degradation of polyphenols. This degradation is attributed to ROS and reactive nitrogen species (RNS) generated during plasma exposure, which can oxidize aromatic rings of polyphenols leading to their fragmentation and reduced antioxidant activity. Furthermore, UV photoionization characteristic of plasma may also contribute to photodegradation of these bioactive compounds, as also noted by Bayati et al.⁷⁵ in their study on plasma-treated fruit juices, underlining the significance of optimizing treatment duration to preserve bioactive compounds. Reactive oxygen and nitrogen species (ROS/RNS) can oxidize the aromatic rings of polyphenols, leading to their fragmentation and formation of degradation products such as di- and carboxylic acids. Additionally, the UV- photoionization characteristic of plasma may be responsible for photodegradation of these compounds, as observed for anthocyanins. A longer cold plasma treatment time (~ 20 min) may therefore cause degradation of polyphenols and loss of antioxidant properties as well as a decrease in the concentration of these bioactive compounds^76,77. On the other hand, the charged particles and UV radiation during CAP formation can cause the breakdown of cell membranes and release of these substances from plant raw materials, increasing their concentration in the final product. This was also confirmed that the polyphenol content equals 12.51 mg GAE/100 ml in pasteurized (63–65 °C, 30 min) guava-flavored whey drinks⁷⁸, while polyphenol content in products treated with N₂-cold plasma (depending on the flow rate of 10, 20, and 30 ml/min for 5, 10, and 15 min) is higher (13.63–14.28 mg GAE/100 ml). In general, the polyphenol concentration in cow’s milk is relatively low compared to the amount typical of raw plant materials or products made from them. However, polyphenols in milk derived from the diet of dairy cattle through rumen fermentation in their digestive system display a potentially higher absorption rate⁷². It is assumed that consuming adequate amounts of low-fat milk reduces the risk of cardiovascular disease, osteoporosis, type 2 diabetes, and even Alzheimer’s disease or colon cancer⁷⁹.

A detailed analysis of the composition of the milk samples (proteins and fats) subjected to the temperature and cold plasma treatment was also performed (Fig. 3S and Table 4S).

Colloidal stability results

The analysis of changes in the TSI (Turbicsan Stability Index) parameter in milk samples exposed to high-temperature and low-temperature plasma allows one to draw conclusions about the colloidal stability of the systems tested (Fig. 2). TSI can comprehensively reflect all instability processes in milk samples, including coalescence, aggregation, and stratification⁸⁰. The lower the TSI value, the greater the stability of the tested system. In this study, as indicated by the present data, fresh cow’s milk that was not subjected to the temperature and cold plasma treatments was the least stable. This may indicate greater aggregation or faster coalescence of particles⁸¹. The other samples tested were characterized by low TSI values, indicating that both the temperature treatment and the cold plasma treatment with nitrogen and air improved the stability of milk. This means that these samples did not show any tendency for significant particle aggregation or coalescence and were stable from a colloidal point of view. Therefore, these treatments appear to prevent efficient particle aggregation or coalescence, maintaining colloidal stability. This observation is in agreement with literature data⁸², which show that cold plasma leads to increased viscosity and a change in the rheological moduli (G’) of milk, which are associated with an increase in colloidal stability due to oxidation and the formation of cross-links between proteins.

Fig. 2.

Changes in the TSI parameter in the tested milk samples. CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

Physicochemical properties results

The analysis of the physicochemical parameters of the cow’s milk (Table 4) showed that the fresh milk had a pH value (6.71) close to neutrality, which is agreement with the literature⁸³and was characterized by a specific conductivity of about 5.72 mS, likely reflecting the presence of dissociable substances in the milk sample studied. The pasteurization process had no effect on pH and the specific conductivity, but the more intensive heat treatment such as sterilization caused a visible decrease in pH by 0.2 units and an increase in specific conductivity, caused by a higher concentration of acid-reacting ionic substances. These compounds could dissociate at higher temperatures. This observed decrease in pH after the heat treatment might be a result of lactose decomposition, which at higher temperatures produces organic acids, e.g. formic acid⁸⁴, or increases the interaction strength of colloidal calcium phosphate, releasing H⁺ ions into the serum, which affects the milk’s pH and conductivity⁸⁵. An additional factor may also be the binding of calcium and phosphates in milk through processing, which causes a decrease in pH ⁸⁶. Similarly, in the cold plasma-treated milk samples, where nitrogen and air were utilized as working gases, a decrease in pH accompanying the increase in specific conductivity was observed, which was associated with the extension of the plasma processing time due to the higher concentration of acidic ionic substances in the tested samples. As far as pH is concerned, the cause of this phenomenon is the partial electrolysis of H₂O molecules and water vapor, which causes an increase in H⁺ cations and a decrease in pH⁸⁷. In turn, the rise in conductivity may be a consequence of accumulation of reactive species during the cold plasma treatment as well as the dissolution of ROS and RNS and the effect of ultraviolet radiation passing through the system⁸⁸. The comparison of the nitrogen and air plasma-treated samples leads to the conclusion that air, which contains oxygen and carbon dioxide in addition to nitrogen, causes greater acidification of the sample and an increase in specific conductivity besides the presence of NO and NO₂ in the plasma, which are also acidification factors. These ROS and RNS can cause the formation of HNO₂ and HNO₃ acids simultaneously with the plasma discharge⁸⁹. The bovine milk sample treated with low-temperature plasma in the presence of air exhibited slightly higher pH values than the sterilized sample, whereas the specific conductivity remained at a comparable level in both cases. These differences were not statistically significant (p ≤ 0.05). These findings suggest that the cold plasma treatment does not markedly alter the acid-base balance or the ionic strength of milk in the tested conditions.

Table 4.

Physicochemical properties of the studied milk samples.

Milk samples	pH	Conductivity [mS]	Z-Ave diameter [nm]
CM	6.71 ± 0.01g	5.072 ± 0.016a	465.70 ± 2.37a
PM	6.69 ± 0.01f	5.085 ± 0.038ab	464.08 ± 6.65a
SM	6.51 ± 0.02a	5.501 ± 0.096e	644.88 ± 10.23b
N5	6.67 ± 0.01e	5.161 ± 0.008abc	464.37 ± 6.95a
N10	6.64 ± 0.01d	5.242 ± 0.010c	462.67 ± 4.93a
N20	6.59 ± 0.01c	5.344 ± 0.028d	463.18 ± 4.88a
A5	6.65 ± 0.01d	5.176 ± 0.042bc	465.08 ± 5.27a
A10	6.60 ± 0.01c	5.252 ± 0.027c	460.00 ± 6.80a
A20	6.54 ± 0.02b	5.512 ± 0.012e	461.72 ± 5.64a

The results are expressed as a mean ± standard deviation. ^{a, b, c…}. - values in columns with different lowercase superscripts differ significantly (p ≤ 0.05). CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

Regarding milk color, which ranges from yellowish-white to nearly pure white, it is influenced by multiple factors, including cow breed and age, milking time, and feed composition. Its characteristic whiteness results from the scattering of visible light by colloidal particles, such as milk fat globules and casein micelles⁹⁰. In this context, the present study (Table 5) revealed significant changes (p ≤ 0.05) in the L* parameter (representing brightness), depending on the preservation method. The CM sample exhibited a lower L* value compared to both the pasteurized (PM) and sterilized (SM) milk samples, between which no significant difference was found. Notably, the treatment with cold atmospheric plasma (CAP) led to an increase in milk brightness, with the highest L* value (p ≤ 0.05) recorded in samples treated with air CAP for 20 min. This enhancement may be attributed to the formation of smaller fat globules, which scatter more effectively²⁵. Furthermore,⁹¹ suggested that the increased brightness may result from enhanced light scattering by casein-calcium colloidal complexes, while other authors linked the reductions in brightness to the shrinkage of casein micelles⁹².

Table 5.

Color parameters of the milk samples.

Milk samples	Color parameters
	L*	a*	b*	ΔE
CM	74.98 ± 0.08a	-1.91 ± 0.02bc	2.53 ± 0.03a	---
PM	76.02 ± 0.04b	-1.87 ± 0.02c	3.56 ± 0.02f	1.46 ± 0.04
SM	75.93 ± 0.04b	0.76 ± 0.04d	8.92 ± 0.03g	6.99 ± 0.02
N5	76.58 ± 0.04c	-2.03 ± 0.02a	2.85 ± 0.05b	1.81 ± 0.06
N10	77.94 ± 0.05g	-2.05 ± 0.03a	2.77 ± 0.05b	2.98 ± 0.08
N20	76.93 ± 0.04d	-1.89 ± 0.03bc	3.40 ± 0.03e	2.44 ± 0.05
A5	77.60 ± 0.03f	-2.07 ± 0.02a	2.86 ± 0.03b	2.64 ± 0.05
A10	77.30 ± 0.03e	-1.95 ± 0.03b	3.15 ± 0.04d	2.40 ± 0.10
A20	78.14 ± 0.04h	-2.04 ± 0.02a	2.98 ± 0.05c	3.20 ± 0.10

The results are expressed as a mean ± standard deviation. ^{a, b, c…}. - values in columns with different lowercase superscripts differ significantly (p ≤ 0.05). CM - control milk sample, PM - pasteurized milk, SM - sterilized milk; nitrogen plasma-treated milk samples: N5–5-min treatment, N10–10-min treatment, N20–20-min treatment; air plasma-treated milk samples: A5–5-min treatment, A10–10-min treatment, A20–20-min treatment.

The sterilized milk was characterized by the lowest value of the a* parameter (0.76) compared to the other samples included in the study. The chromaticity coefficient for the control, pasteurized, and low-temperature preserved milk ranged from − 1.87 to -2.07. Heating the milk at a high temperature most likely caused the destabilization of casein micelles, which significantly (p ≤ 0.05) increased the a* value. The oxidation of milk proteins occurs under the influence of cold plasma, causing only a slight yellowing effect⁹³, which was also detected in our study. A characteristic feature of the color of the sterilized milk was the increase in yellow hue in comparison to the control milk. Changes in the average values of the b* parameter in the other preserved samples were not as noticeable. Heating milk can affect the formation of colored melanoidins responsible for a change in the b* chromaticity coefficient of the milk color towards brown. The basis of these changes is the Maillard reaction, during which a series of simultaneous chemical reactions occur. It begins with the binding of the lactose molecule’s aldehyde group to the ε-amino group of lysine residues found in various milk proteins and, as a result, is responsible not only for the taste and smell of products but also for their color.

The delta E coefficient (ΔE) is a quantity that aligns with the human visual perception of the variation between two observed colors. The calculated value of the parameter⁹⁴in the range of 1 to 2 indicates that the difference in color between the CM samples and the PM and N5 samples will be noticeable only to the eye of an experienced observer. ΔE in the range of 2 to 3.5 for the samples processed with cold plasma (10 and 20 min) with the nitrogen working gas and air (5, 10, and 20 min) indicates a difference that is obvious to any observer. In turn, a value of 6.99 expresses a large deviation, i.e. the difference between the control milk sample and the sterilized one is clearly perceived as two different colors. ²⁹ when processing cow’s milk, suggested that an atmospheric corona discharge lasting up to 15 min could result in ΔE values within the range of an imperceptible color change, which is also in line with the data obtained by⁹¹ for sheep’s milk. The authors did not show a clear trend in the ΔE color changes when increasing the DBD cold plasma treatment time from 30 to 180 and up to 300 s. However, the maximum difference in color (1.12) was observed after pasteurization (65 °C ± 2 °C for 30 min). Meanwhile, Manoharan⁶⁶observed a change in the ΔE value in cow’s milk treated with air cold plasma in a continuous plasma system based on a DBD reactor when they changed the flow rate from 3 ml/min to 6 ml/min. The procedures performed by⁸⁷ in a DBD reactor at variable voltages of 0, 40, 50, 60, 70, and 80 V for 120 s and for 0, 15, 30, 60, 90, and 120 s at a voltage of 70 V also resulted in a change in the above-mentioned parameter.

Additional observations can be related to the chemical interactions between plasma-generated reactive species and milk components. Cold plasma produces reactive oxygen and nitrogen species (ROS, RNS), such as hydroxyl radicals, ozone, and nitric oxides, which may react with proteins, amino acids, and lipids⁹⁵. These reactions can induce mild oxidation or cross-linking of amino acid residues, formation of carbonyl groups, or cleavage of disulfide bridges, resulting in subtle structural modifications of milk proteins^96,97. Such changes can influence particle size distribution and light scattering, explaining minor variations in color parameters. Water molecules in milk, also participate in plasma-induced reactions, acting as a source of hydroxyl radicals and contributing to the slight acidification and conductivity increase observed after treatment. Although plasma exposure can potentially oxidize minerals or degrade vitamins, our results indicate that, under the applied conditions (≤ 36 °C, 5–20 min), these effects were limited, as the concentrations of vitamins and minerals remained comparable to or higher than those in thermally treated samples.

Particle size measurements results

The size of the particles in the milk emulsion has a major impact on its taste, consistency, and stability⁹⁸. It should be noted, however, that these particle size changes are influenced by multiple factors, including cow breed, feed, and season⁹⁹. From a sensory perspective, particle size plays a crucial role in determining the sensory quality and physical characteristics of milk¹⁰⁰. Smaller particle sizes contribute to a smoother and more uniform texture. On the other hand, larger particles may lead to a coarser or uneven mouthfeel¹⁰¹. Smaller particles enhance colloidal stability, thus minimizing creaming and sedimentation over time. Finer emulsions allow more efficient interaction between the fat phase and flavor compounds, enhancing the delivery of aroma and taste molecules to sensory receptors. Smaller particles may better encapsulate volatile compounds, leading to prolonged flavor retention, while larger aggregates can trap flavors unevenly or release them prematurely¹⁰². In fresh raw cow’s milk, a bimodal distribution of particle sizes is most often observed: a peak of 0.20 μm, representing casein micelles, and a second peak of 3.7 μm, corresponding to fat globules⁹⁹. The size of fat globules in milk is usually between 0.10 μm and 15 μm¹⁰³. The results of the particle size measurements show that fat droplets dispersed in the water had an average size of around 476 nm, which is in agreement with the literature¹⁰⁴. Pasteurization causes only slight changes in particle size. The sterilized milk sample was quite different, as it was characterized by much larger average particle sizes, with an average of 645 nm. This behavior is most likely caused by the coalescence of the fat droplets that takes place under high temperature conditions. This effect is a consequence of protein oxidation at high temperature, which causes peptide chain scission, protein aggregation, and cross-linking¹⁰⁵. Moreover, higher temperature is responsible for an increase in the kinetic energy of the particles, resulting in faster particle motion, which can lead to more frequent collisions of fat particles, and consequently, to coalescence. This observation is consistent with the microscopic analysis performed (Fig. 3), which also confirms the presence of larger fat particles in the sterilized milk. On the other hand, according to the literature, the effect of plasma processing on the milk droplet size may differ. Larger droplets after cold plasma treatment are regarded as a consequence of the presence of ROS and RNS in plasma-treated milk, which cause mild oxidation of proteins, an increase in the number of carbonyl groups, surface lipophilicity and a reduction of free SH groups⁹⁹. Conversely, there are also some reports⁹²that plasma processing can lead to a reduced size of droplets in comparison to the control sample. Our findings showed no significant differences between the CM sample and the plasma processed ones. The present data also indicate that there is no influence of the plasma treatment time on the milk droplet sizes, which is consistent with literature findings⁹². Moreover, the comparison of the results obtained for the plasma-treated milk samples with the use of different working gases showed no significant changes between the samples.

Fig. 3.

Microscopic images of the studied milk samples: (a) CM - control milk sample; (b) PM - pasteurized milk; (c) SM - sterilized milk; (d) N5 - milk treated with nitrogen plasma for 5 min; (e) N10 - milk treated with nitrogen plasma for 10 min; (f) N20 - milk treated with nitrogen plasma for 20 min; (g) A5 - milk treated with air plasma for 5 min; (h) A10 - milk treated with air plasma for 10 min; (i) A20 - milk treated with air plasma for 20 min.

Conclusions

This study compared traditional thermal methods of milk sanitization with cold plasma treatment. The results clearly indicate that cold plasma treatment represents a viable alternative for milk hygienization.

Toxicological and Microbiological safety The studies using Danio rerio larvae demonstrated that plasma-treated milk is a safer alternative to standard thermal processing, as indicated by higher survival rates of zebrafish. The optimal plasma exposure time ensuring both high microbiological safety and low toxicity was 10 min for nitrogen or air plasma treatments. Although cold plasma was less effective than thermal sterilization in achieving complete bacterial elimination, it still provided microbial inactivation within acceptable safety standards, without the need for high temperatures.

Nutritional value and chemical composition Cold plasma-treated milk retained higher concentrations of B vitamins, minerals, and polyphenols compared with thermally hygienized samples. This preservation of bioactive compounds highlights the advantage of non-thermal processing in maintaining the nutritional quality of milk, an aspect crucial for both consumer health and product functionality.

Colloidal and physicochemical stability The colloidal stability of cold plasma treated milk was similar to that of heat-treated samples but higher than that of untreated milk. Moreover, cold plasma treatment caused no significant changes in basic physicochemical parameters (pH, particle diameter, color), in contrast to the pronounced alterations observed after sterilization.

Practical implications These findings, taken together, support the potential of cold plasma as an energy-efficient, environmentally friendly technology for milk preservation. Further research should focus on optimizing plasma processing conditions to ensure industrial scalability while preserving both the microbiological safety and full nutritional value of milk.

Future studies The application of cold atmospheric plasma (CAP) in milk processing has shown promise in inducing structural changes in milk proteins, as evidenced by FTIR spectral analysis. These changes may involve partial degradation or cross-linking of amino acid residues. Nevertheless, the current study did not employ advanced analytical techniques, such as high-performance liquid chromatography (HPLC) or amino acid profiling using dedicated analyzers, to confirm and quantify these modifications. Considering the nutritional importance of proteins and amino acids in milk, future research should incorporate such methods to provide a more comprehensive understanding of CAP-induced effects on milk quality and composition. Future studies will also focus on the quantitative determination of nitrate and nitrite levels in plasma-treated milk, particularly when nitrogen is used as the working gas, to confirm the absence of potentially harmful residues. In addition, sensory analysis should be conducted to evaluate the flavor, aroma, and overall acceptability of plasma-treated milk, ensuring that the process maintains both safety and consumer appeal.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

E. G. Conceptualization, Methodology, Software, Writing—original draft, Writing—review & editing, Supervision, Project administration, M. K. Methodology, Writing—original draft, Statistical analysis, B. B. Conceptualization, Methodology, Analytical part, Writing—original draft, Ł. K. Methodology, Software, Writing—original draft, P. T. Investigation, Visualization, Writing—review & editing, D. Z. Methodology, Data analysis, M. K. Methodology, Data analysis, J. P. Methodology, Conceptualization, Writing—review & editing, K. G. Analytical part, Visualization, A. S-W. Methodology, Conceptualization, Data analysis, Analytical part, Writing—original draft, Writing—review & editing.

Funding

The project was financed by Union of Lublin Universities in the framework of the program “INTERPROJECT” edition III, (principal investigator: E. Grządka)

Data availability

Data are contained within the article as well as in the Supplementary material.

Declarations

Competing interests

The authors declare no competing interests.