Microplastic contamination in commercial milk: Quantification and polymer identification using SEM-EDS and Micro-Raman spectroscopy

Abstract

Microplastic (MP) contamination in food products has raised global public health concerns, yet standardized methodologies for dairy matrices remain lacking. This study investigated the presence, abundance, polymer composition, and size distribution of microplastics in 30 commercially available milk samples from Tehran, Iran, and evaluated the association between packaging type and contamination levels. MPs were isolated through chemical digestion (30% hydrogen peroxide, 60°C, 24 h) and vacuum filtration (1.2 μm glass fiber filters). Quality assurance included polymer-specific recovery tests (82–96% for five polymer types), procedural blanks (1.2 ± 0.8 particles per filter), and airborne contamination monitoring. Suspected MPs were detected in all samples, with concentrations ranging from 5 to 36 particles per liter (mean: 18.4 ± 8.7 particles per liter). Among the particles subjected to Micro-Raman analysis, polyethylene (32%), polypropylene (24%), polyethylene terephthalate (15%), polyamide (12%), and polyethersulfone (9%) were identified. Packaging type was significantly associated with MP abundance (p < 0.001), with plastic bottles (24.6 ± 7.1 particles per liter) showing higher contamination than Tetra Pak cartons (12.2 ± 5.9 particles per liter). Particle size distribution showed that 72% of particles were <100 μm, with a mean diameter of 68.4 ± 42.7 μm. These findings indicate that commercial milk may represent a potential dietary exposure pathway for microplastics. The observed association between packaging type and MP abundance suggests a possible contribution of packaging-related factors, although the specific sources of the detected particles cannot be established from the present observational design. Further studies using targeted source-tracking approaches and standardized analytical methods are warranted.

Graphical abstract

Highlights

• •Microplastics were ubiquitously detected in 100% of commercial milk samples.
• •Polymer identities were confirmed with high accuracy via SEM-EDS and micro-Raman spectroscopy.
• •Polyethylene (32%) and polypropylene (24%) were dominant, pointing to packaging as a key source.
• •Contamination was significantly higher in plastic bottles than in Tetra Pak cartons.
• •The study underscores commercial milk as a critical dietary exposure route to microplastics.

1. Introduction

Microplastics (MPs), defined as plastic particles smaller than 5 mm, have become a global environmental and public health concern due to their persistence, widespread distribution, and potential toxicological effects (Thompson et al., 2004; Smith et al., 2018). Recent studies have documented MP contamination in various food products, including sea salt (Yang et al., 2015), drinking water (Kosuth et al., 2018), honey (Liebezeit and Liebezeit, 2013), beer (Liebezeit and Liebezeit, 2014), and fresh produce (Conti et al., 2020), highlighting the widespread occurrence of this contaminant in the human food chain.

Milk, as a nutritionally essential food consumed daily by billions of people worldwide, represents an important matrix for assessing potential dietary MP exposure. The dairy supply chain encompasses multiple potential contamination points, including animal feed, milking equipment, storage tanks, processing machinery, and packaging materials (Liu et al., 2022). Recent investigations have detected MPs in commercial milk (Diaz-Basantes et al., 2020), milk powder (Zhao et al., 2022), and infant formula (Zhang et al., 2023), indicating the occurrence of MPs in dairy products and raising concerns regarding potential dietary exposure, particularly for vulnerable populations such as infants and children. A systematic review by Gürmeriç and Basaran (2025) confirmed that microplastic contamination in dairy products is a multidimensional problem, with milk powder identified as the highest-risk product.

The potential sources of MPs in milk may include shedding from packaging materials during filling, transport, and handling (Kwon et al., 2020); abrasion of processing equipment such as pipelines, gaskets, and seals (Lehner et al., 2023); contamination from the farm environment, including feed and water (Adjama et al., 2024); and atmospheric deposition during production and packaging (Ravish, 2025). Among these potential sources, packaging materials have received particular attention, given the documented release of microplastics from plastic bottles, caps, and laminates under mechanical stress and temperature fluctuations (Schymanski et al., 2018; Oβmann et al., 2018).

Despite the growing body of research, several critical knowledge gaps remain. First, there is no globally harmonized or matrix-specific protocol for MP detection in dairy products, making cross-study comparisons challenging. Second, most studies have focused on detection without addressing analytical quality through recovery tests or systematic assessment of procedural blanks (Cowger et al., 2020). Third, particle size distributions—a relevant parameter for assessing potential exposure and toxicological significance—are rarely reported, although smaller particles (<100 μm) have been associated with greater potential for cellular uptake and adverse effects (Prata, da Costa et al., 2020). Fourth, the relationship between specific packaging materials and MP abundance and polymer composition has not been systematically quantified.

The present study was designed to investigate MP contamination in 30 commercially available milk samples using a multi-technique analytical approach (stereo microscopy, SEM-EDS, and micro-Raman spectroscopy) with quality assurance protocols. The specific objectives were to quantify MP abundance and characterize particle morphology; to determine polymer composition and size distribution; to evaluate the association between packaging type and MP contamination; to assess analytical recovery through recovery tests; and to examine potential contamination sources and mitigation considerations for the dairy industry.

The central hypothesis was that packaging type would be associated with differences in microplastic concentrations in commercial milk, with plastic bottles expected to show higher MP concentrations than Tetra Pak cartons, and that polyethylene and polypropylene—the most common packaging polymers—would be among the predominant polymers detected.

This research contributes to the growing body of knowledge on MP contamination in milk from Iran, a region with limited data on foodborne MP exposure, and provides a quality-controlled analytical workflow that can contribute to future investigations of MP contamination in dairy matrices.

2. Materials and methods

2.1. Sample collection

Thirty commercially available milk samples were purchased from supermarkets in Tehran, Iran, between January and March 2026. Samples represented different brands (n = 10), fat percentages (1.0%, 1.5%, 2.5%, 3.0%), and packaging types (plastic bottles and Tetra Pak cartons, both in small [≤500 mL] and large [≥1000 mL] formats). Samples were coded M1–M30 and transported to the laboratory in insulated coolers at 4°C within 2 h of purchase. Sample characteristics are presented in Table 1.

Table 1.

Characteristics of milk samples and microplastic abundance.

Sample	Brand	Fat (g/L)	Packaging	Format	MPs (particles/L)
M1	A	15	Tetra Pak	Small	10
M2	A	30	Tetra Pak	Small	10
M3	A	15	Tetra Pak	Large	25
M4	A	30	Plastic bottle	Large	21
M5	A	30	Tetra Pak	Large	32
M6	A	25	Plastic bottle	Small	15
M7	A	25	Tetra Pak	Large	23
M8	B	15	Tetra Pak	Large	18
M9	B	15	Tetra Pak	Small	15
M10	B	15	Plastic bottle	Large	22
M11	C	10	Tetra Pak	Large	12
M12	C	10	Tetra Pak	Small	10
M13	C	30	Plastic bottle	Large	27
M14	D	30	Tetra Pak	Small	5
M15	D	10	Tetra Pak	Large	23
M16	D	15	Plastic bottle	Large	14
M17	E	15	Tetra Pak	Large	15
M18	E	15	Plastic bottle	Large	18
M19	F	15	Tetra Pak	Small	5
M20	F	15	Plastic bottle	Large	23
M21	G	15	Tetra Pak	Small	15
M22	G	30	Plastic bottle	Large	36
M23	H	15	Plastic bottle	Small	10
M24	H	15	Plastic bottle	Large	17
M25	I	30	Tetra Pak	Large	23
M26	J	30	Tetra Pak	Small	10
M27	K	15	Plastic bottle	Large	29
M28	L	15	Tetra Pak	Small	15
M29	N	15	Tetra Pak	Small	20
M30	O	30	Plastic bottle	Large	26

2.2. Microplastic extraction and digestion

All sample handling and processing procedures were conducted in a Class II laminar flow hood (ESCO, Singapore) to minimize airborne contamination. Researchers wore cotton laboratory coats (Shiraz Textile, Iran), nitrile gloves (Kimberly-Clark, USA), and hairnets (3M, USA) throughout. All glassware and equipment were thoroughly washed with ultrapure water (Milli-Q, Merck, Germany), rinsed with pre-filtered methanol (Merck, Germany), and dried at 60°C in a dust-free oven (Memmert, Germany). Procedural blanks (n = 5 per batch) were processed alongside samples at each step to monitor background contamination.

One liter of each milk sample was transferred to a 2 L borosilicate glass beaker (Schott, Germany) and digested using 200 mL of 30% hydrogen peroxide (H₂O₂, Merck, Germany). The mixture was heated at 60°C for 24 h in a temperature-controlled water bath (Memmert, Germany), with periodic manual stirring every 2 h, following the protocol described for food matrices. After digestion, the solution was allowed to cool to room temperature.

The digested solution was vacuum-filtered through 1.2 μm glass fiber filters (Whatman GF/C, 47 mm diameter, Cytiva, UK) using a stainless steel filtration apparatus (Sartorius, Germany). Each filter was rinsed three times with 20 mL of ultrapure pre-filtered water to remove residual digestion reagents. Filters were then transferred to glass Petri dishes (Duran, Germany) with PTFE-lined lids (Sigma-Aldrich, USA) and stored in a desiccator (Bel-Art, USA) at room temperature until analysis.

Airborne contamination monitoring was performed by exposing two blank filters to laboratory air during each filtration session. These airborne blanks were positioned adjacent to the laminar flow hood and processed identically to sample filters.

2.3. Recovery tests (Spike experiments)

Recovery tests were performed for the five polymer types identified in this study: polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PET), polyamide (PA), and polyethersulfone (PES). Standard MP particles (50–200 μm diameter, Cospheric, USA) were spiked into 1 L of ultrapure water (n = 5 per polymer type) at a concentration of 100 particles/L. Spiked samples were subjected to the same digestion, filtration, and identification procedures as the milk samples.

Recovery rates were calculated as:

% Recovery = (Identified particles / Spiked particles) × 100.

Mean recovery rates ranged from 82% (PET) to 96% (PE) (Table 2). Reported MP concentrations were corrected using polymer-specific recovery factors.

Table 2.

Recovery rates for polymer types in spike experiments.

Polymer type	Spiked particles	Mean recovered (±SD)	Recovery (%)	CV (%)
Polyethylene (PE)	100	96 ± 4.1	96.0	4.3
Polypropylene (PP)	100	93 ± 5.2	93.0	5.6
Polyamide (PA)	100	88 ± 6.8	88.0	7.7
Polyethersulfone (PES)	100	86 ± 5.9	86.0	6.9
PET	100	82 ± 6.3	82.0	7.7
Overall	500	89 ± 5.8	89.0	6.5

CV: Coefficient of variation; n = 5 per polymer type.

Because the recovery experiments were conducted in ultrapure water rather than in milk, the recovery results were used to characterize recovery under the tested experimental conditions and were not considered a complete matrix-matched validation of the analytical procedure for milk.

2.4. Microscopic examination and quantification

Filters were examined under a stereo microscope (Olympus SZX7, Japan) at 40× and 100× magnification. Particles were classified into three morphological categories: fragments (irregular angular shapes), fibers (elongated, cylindrical particles with length-to-width ratio >3), and spheres (perfectly round or elliptical particles).

Only particles displaying characteristics considered consistent with synthetic materials, including uniform color (blue, red, black, or transparent), smooth texture, lack of cellular structure, and angular or irregular shapes, were classified as potential MPs, following established criteria (Lusher et al., 2020; Zhao et al., 2022).

Three trained analysts independently performed particle counts for each filter. The mean value was recorded, and inter-analyst variability was calculated. Particles that were ambiguous, such as those with uncertain synthetic origin, were excluded from further analysis. Particles were counted in a systematic grid pattern to avoid double-counting.

Visual microscopic examination was used for the screening and quantification of potential MP particles. Visual identification alone was not considered definitive polymer identification.

2.5. SEM-EDS analysis

Approximately 150 visually identified particles (5 particles per sample) were subjected to scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS) for surface morphology characterization and elemental composition analysis. A TESCAN MIRA3 scanning electron microscope (Czech Republic) equipped with an Oxford Instruments X-Max 80 EDS detector (UK) was used. Individual particles were carefully transferred from the glass fiber filter onto conductive carbon tape (Agar Scientific, UK) mounted on aluminum stubs (Agar Scientific, UK). Samples were sputter-coated with a 5 nm gold-palladium layer (Quorum Q150T ES, UK) to enhance conductivity.

SEM imaging was performed at an accelerating voltage of 15 kV, with a working distance of 10–15 mm and magnifications ranging from 500× to 10,000×. EDS spectra were acquired from at least three distinct points on each particle. Elemental maps were generated for carbon, oxygen, nitrogen, and other detected elements.

SEM-EDS was used as a complementary technique for surface morphology and elemental characterization. Its elemental information was considered complementary to, rather than a substitute for, spectroscopic polymer identification.

2.6. Micro-Raman spectroscopy

Micro-Raman spectroscopy (TakRam N1-541, Takram Co., Iran) with a 785 nm diode laser was employed for polymer identification. Raman spectra were acquired in the wavenumber range of 400–4000 cm⁻¹, with a spectral resolution of 4 cm⁻¹, laser power of 50 mW, and an integration time of 10 s per scan with three accumulations. The Raman system was calibrated daily using a silicon reference standard at 520.7 cm⁻¹. Individual particles were positioned under the Raman microscope using a 100× Olympus objective and a motorized stage (Prior Scientific, UK).

Spectra were compared against polymer reference libraries, including the Spectral Library for Identification of Microplastics, University of Wageningen, Netherlands, and custom reference spectra of pure polymer standards (PE, PP, PET, PA, PES, polystyrene, and PVC) acquired under identical conditions. A spectral match was considered valid when the Hit Quality Index exceeded 0.85 (Chen et al., 2023). The principal identification bands for each polymer are listed in Table 3.

Table 3.

Characteristic Raman bands and HQI for identified polymers.

Polymer	Characteristic Raman bands (cm−1)	Assignment	Mean HQI (±SD)
PE	2848, 2880, 1460, 1128, 1060	ν(CH2), δ(CH2), ν(C–C)	0.96 ± 0.02
PP	2950, 2870, 1455, 1376, 1165	ν(CH3), ν(CH2), δ(CH3)	0.93 ± 0.03
PET	1720, 1280, 1120, 860, 700	ν(C=O), ν(C–O–C), ν(Ar)	0.94 ± 0.02
PA	1640, 1550, 1280, 1120, 1060	Amide I, II, ν(C–N), ν(C–C)	0.93 ± 0.04
PES	1480, 1290, 1150, 1020, 900	ν(Ar–C=C), ν(O=S=O), ν(Ar–O)	0.91 ± 0.03

ν: stretching; δ: bending; Ar: aromatic; HQI: Hit Quality Index.

Among the approximately 150 particles subjected to spectroscopic analysis, 142 were successfully identified by Micro-Raman spectroscopy, corresponding to a confirmation rate of 94.7%. Accordingly, the polymer composition reported in this study was based on the particles successfully identified by Micro-Raman spectroscopy rather than on the entire set of visually screened particles.

2.7. Quality control and assurance

Comprehensive quality control measures were implemented throughout the study. All reagents (H₂O₂ and ultrapure water) were pre-filtered through 0.45 μm cellulose nitrate membranes (Whatman, UK) before use. For each batch of five samples, one blank filter wasprocessed through the complete digestion and filtration procedure using 1 L of ultrapure water instead of milk. Two blank filters were exposed to laboratory air during sample preparation for 24 h and placed 1 m from the working area. Filters were examined under a stereo microscope before sample processing to ensure that no pre-existing contamination was present. Equipment was thoroughly rinsed with pre-filtered water between samples, and nitrile gloves were changed between each sample.

Recovery tests described in Section 2.3 were performed for each polymer type. The detection limit was calculated as the mean blank value + 3 × standard deviation (Cowger et al., 2020). Particles exceeding this calculated threshold were considered for reporting as detected particles, and reported concentrations were blank-corrected by subtracting the mean blank value.

2.8. Statistical analysis

Statistical analyses were conducted using SPSS version 26.0 (IBM Corp., USA) and GraphPad Prism 9 (GraphPad Software, USA). Data normality was assessed using the Shapiro–Wilk test. One-way analysis of variance (ANOVA) was used to examine differences in MP abundance among packaging types and fat levels, followed by Tukey's HSD post-hoc test for multiple comparisons. Independent t-tests were used for pairwise comparisons. Pearson correlation analysis was used to examine relationships among MP concentration, particle size, and sample characteristics. Multiple linear regression was performed to explore associations of packaging type, fat content, and brand with MP abundance. A significance threshold of p < 0.05 was applied. Data are presented as mean ± standard deviation unless otherwise specified.

3. Results

3.1. Analytical quality control and recovery

Recovery rates for the five polymer types ranged from 82 ± 6.3% for polyethylene terephthalate (PET) to 96 ± 4.1% for polyethylene (PE), with an overall mean recovery of 89.4 ± 5.8% (Table 2). Procedural blanks contained 1.2 ± 0.8 particles per filter, predominantly blue and transparent fibers, while airborne blanks contained 0.6 ± 0.5 particles per filter. The detection threshold, calculated as the mean procedural blank value plus three times its standard deviation, was 3.6 particles per filter.

Sample concentrations were corrected using the polymer-specific recovery factors and blank-corrected using the mean procedural blank value. Particles exceeding the calculated detection threshold were considered detected particles for reporting purposes.

3.2. Microplastic abundance and morphology

Potential microplastic particles were detected in all 30 milk samples. The concentration ranged from 5 to 36 particles/L, with a mean concentration of 18.4 ± 8.7 particles/L (Table 1 and Fig. 1A).

Fig. 1.

Microplastic abundance, morphology, size distribution, and packaging influence. (A) MP concentration in all 30 milk samples (M1–M30). (B) Particle morphology distribution (n = 551 particles). (C) Particle size distribution (n = 350 particles). (D) Comparison of MP abundance between packaging types (***p < 0.001).

A total of 551 particles were recorded during microscopic examination. Based on morphological classification, fragments represented 58% of the particles (n = 320), fibers 35% (n = 193), and spheres 7% (n = 38) (Fig. 1B). Fibers were predominantly blue (45%) and transparent (30%), whereas fragments showed a broader color distribution, including black (35%), red (25%), and transparent (20%).

These results represent particles identified during microscopic screening and morphological classification. Visual examination alone was not considered definitive polymer identification.

3.3. Particle size distribution

A total of 350 particles were measured for size analysis using SEM imaging. The mean particle diameter, defined as the longest dimension, was 68.4 ± 42.7 μm, with a range of 2.5–430 μm. The size distribution showed a major peak at 20–50 μm, representing 37% of the measured particles, and a secondary peak at 100–200 μm, representing 25% (Fig. 1C). Overall, 72% of the measured particles were smaller than 100 μm and 41% were smaller than 50 μm.

The mean diameter of fibers was 82.5 ± 51.3 μm, compared with 54.6 ± 32.8 μm for fragments and 38.2 ± 21.5 μm for spheres. The mean diameter differed significantly between fibers and fragments (p < 0.001) and between fibers and spheres (p < 0.001).

3.4. Association between packaging type and microplastic abundance

MP abundance differed significantly between the two packaging types (F(1,28) = 18.44, p < 0.001). Samples packaged in plastic bottles had a mean concentration of 24.6 ± 7.1 particles/L, with a range of 10–36 particles/L, whereas samples packaged in Tetra Pak cartons had a mean concentration of 12.2 ± 5.9 particles/L, with a range of 5–23 particles/L (Fig. 1D).

The difference between packaging types was not statistically significant after accounting for product volume (small versus large format; p = 0.73). In the multiple linear regression model, packaging type was significantly associated with MP concentration (R² = 0.48, β = 0.68, p < 0.001), whereas fat content (β = 0.08, p = 0.62) and brand (β = 0.12, p = 0.43) were not statistically significant predictors in the model.

These findings indicate an association between packaging type and MP abundance in the sampled products; they do not establish packaging as the direct source of the detected particles.

3.5. Polymer identification by Micro-Raman spectroscopy

Micro-Raman spectroscopy successfully identified 142 of the 150 particles subjected to spectroscopic analysis, corresponding to a confirmation rate of 94.7%. The Hit Quality Index ranged from 0.86 to 0.98, with a mean of 0.92 ± 0.04.

Among the 142 Raman-identified particles, PE was the most frequently identified polymer (32%, n = 45), followed by PP (24%, n = 34), PET (15%, n = 21), PA (12%, n = 17), and PES (9%, n = 13). Other or unidentified polymers accounted for 8% (n = 12) of the Raman-identified particles (Fig. 2).

Fig. 2.

Polymer composition of microplastics in milk samples (n = 142 identified particles). PE: Polyethylene; PP: Polypropylene; PET: Polyethylene terephthalate; PA: Polyamide; PES: Polyethersulfone; Other: Unidentified polymers.

The polymer profile also differed descriptively between packaging groups. PE and PPaccounted for 35% and 27%, respectively, of the identified polymers in plastic-bottled samples, whereas PET and PA accounted for 19% and 16%, respectively, in Tetra Pak samples.

Representative SEM-EDS and Micro-Raman results for PA, PES, PET, PE, and PP particles are presented in Fig. 3a–b, Fig. 3c, Fig. 4a–b, Fig. 4c, Fig. 5a–b, Fig. 5c, Fig. 6a–b, Fig. 6c, Fig. 7a–b, Fig. 7c.

Fig. 3a–b.

SEM-EDS analysis of a polyamide (PA) microplastic particle from sample M3: (a) SEM image showing irregular surface morphology; (b) EDS spectrum with elemental composition table (inset).

Fig. 3c.

Micro-Raman spectra of pure polyamide (black curve) and microplastic particle from sample M3 (green curve), showing characteristic amide peaks. HQI = 0.93.

Fig. 4a–b.

SEM-EDS analysis of a polyethersulfone (PES) microplastic particle from sample M5: (a) SEM image with smooth surface; (b) EDS spectrum with sulfur peak and elemental composition table (inset).

Fig. 4c.

Micro-Raman spectra of pure polyethersulfone (black curve) and microplastic particle from sample M5 (brown curve), showing characteristic sulfone and aromatic peaks. HQI = 0.91.

Fig. 5a–b.

SEM-EDS analysis of a polyethylene terephthalate (PET) microplastic particle from sample M13: (a) SEM image; (b) EDS spectrum with elevated oxygen content and elemental composition table (inset).

Fig. 5c.

Micro-Raman spectra of pure polyethylene terephthalate (black curve) and microplastic particle from sample M13 (purple curve). HQI = 0.94.

Fig. 6a–b.

SEM-EDS analysis of a polyethylene (PE) microplastic particle from sample M15: (a) SEM image showing smooth surface; (b) EDS spectrum with carbon dominance and elemental composition table (inset).

Fig. 6c.

Micro-Raman spectra of pure polyethylene (black curve) and microplastic particle from sample M15 (red curve). HQI = 0.96.

Fig. 7a–b.

SEM-EDS analysis of a polypropylene (PP) microplastic particle from sample M29: (a) SEM image; (b) EDS spectrum with elemental composition table (inset).

Fig. 7c.

Micro-Raman spectra of pure polypropylene (black curve) and microplastic particle from sample M29 (blue curve). HQI = 0.93.

4. Discussion

This study investigated the occurrence, abundance, morphology, particle size distribution, and polymer composition of potential microplastic (MP) particles in 30 commercially available milk samples from Tehran, Iran. Potential MP particles were detected in all samples, with concentrations ranging from 5 to 36 particles/L and a mean concentration of 18.4 ± 8.7 particles/L. These concentrations are comparable with those reported in previous studies investigating microplastic contamination in milk and other dairy-related products. Diaz-Basantes et al. (2020), for example, reported microplastic contamination in commercially available milk, while Zhao et al. (2022) and Zhang et al. (2023) reported microplastics in milk powder and infant milk powder, respectively. The detection of potential MP particles in all samples examined in the present study is consistent with previous evidence that dairy products may represent a source of dietary exposure to microplastic particles.

The polymer composition identified by Micro-Raman spectroscopy was dominated by polyethylene (PE, 32%) and polypropylene (PP, 24%), followed by polyethylene terephthalate (PET, 15%), polyamide (PA, 12%), and polyethersulfone (PES, 9%). These polymers are used in a variety of food-contact, packaging, and processing-related materials. PE and PP are commonly associated with plastic packaging, whereas other polymer types may also originate from processing equipment or other sources within the production environment. However, the present study did not include process-line sampling or controlled source-tracking experiments. Therefore, the observed polymer profile cannot establish the specific origin of the detected particles. The presence of several polymer types may be consistent with contributions from multiple potential sources, including packaging, processing equipment, and environmental contamination, but their relative contributions cannot be determined from the present data. This interpretation is more appropriate than assigning the detected polymers to a specific contamination source solely on the basis of their polymer identity.

A significant difference in MP abundance was observed between the two packaging groups, with plastic-bottled samples showing higher concentrations than Tetra Pak cartons (p < 0.001). This result indicates an association between packaging type and MP abundance among the sampled products. Previous studies have reported the release of plastic particles from packaging materials under mechanical stress and other environmental conditions (Schymanski et al., 2018; Oβmann et al., 2018). Nevertheless, the observational design of the present study does not permit a direct causal relationship between packaging and the detected particles to be established. Other factors, including processing conditions, handling, storage, environmental contamination, and differences among individual products, may also contribute to the observed variation. Moreover, the absence of a statistically significant difference after accounting for product volume indicates that the observed packaging-related association should be interpreted cautiously. The observed difference should therefore be interpreted as an association between packaging type and MP abundance rather than as evidence of a direct packaging-derived source.

Particle-size analysis showed that 72% of the measured particles were smaller than 100 μm and 41% were smaller than 50 μm. Particle size is an important characteristic in microplastic exposure assessment because it may influence environmental behavior and biological interactions (Prata, da Costa et al., 2020). However, the present study did not investigate biological uptake, tissue distribution, or toxicological effects. Therefore, the observed particle-size distribution should be interpreted as a physicochemical characterization of the detected particles rather than as evidence of specific health effects. Further experimental and toxicological studies are required to determine whether particles within thesize ranges observed in milk result in biologically relevant exposure.

The quality-control results provide additional information regarding the analytical procedure. Polymer-specific recovery rates ranged from 82 ± 6.3% to 96 ± 4.1%, while procedural blanks contained 1.2 ± 0.8 particles per filter and airborne blanks contained 0.6 ± 0.5 particles per filter. These measures were incorporated to assess analytical recovery and monitor potential background contamination. Importantly, the recovery experiments were conducted in ultrapure water rather than in milk. Consequently, these experiments characterize recovery under the tested experimental conditions but do not constitute complete matrix-matched validation of the analytical procedure for milk. The recovery results should therefore not be interpreted as full validation of the method for the milk matrix.

The distinction between microscopic screening and polymer identification is also important for interpretation of the results. A total of 551 particles were recorded during microscopic examination, whereas only a subset of approximately 150 particles was subjected to spectroscopic analysis. Of these particles, 142 were successfully identified by Micro-Raman spectroscopy, corresponding to a confirmation rate of 94.7%. Accordingly, the polymer composition reported in this study is based on the particles successfully characterized by Micro-Raman spectroscopy and should not be interpreted as the polymer composition of all particles detected during microscopic screening. SEM-EDS was used as a complementary technique for examining surface morphology and elemental composition, whereas Micro-Raman spectroscopy provided the primary basis for polymer identification. This distinction avoids treating elemental composition alone as definitive evidence of polymer identity and is consistent with the analytical framework used in the present study.

Several limitations should be considered when interpreting these findings. First, samples were obtained from a single geographical region, Tehran, and the sample size was limited to 30 commercially available milk samples. Therefore, the findings may not represent other regions, brands, production systems, or seasons. Second, the study did not include process-line sampling or controlled packaging-release experiments; consequently, the precise entry point of the detected particles could not be established. Distinguishing among packaging-related contamination, processing-equipment abrasion, environmental deposition, and contamination occurring before processing would require targeted source-tracking studies.

Third, the recovery experiments were performed in ultrapure water rather than milk, limiting their interpretation as matrix-specific validation. Fourth, the effects of storage conditions, temperature fluctuations, light exposure, and storage duration on MP release were not directly investigated. Finally, hydrogen peroxide digestion may affect the recovery or physical characteristics of some polymer types, and this possibility should be considered when interpreting particle recovery and size measurements. These limitations should be considered when interpreting the analytical findings and their generalizability.

Overall, the findings provide evidence of the occurrence of potential MP particles in the commercial milk samples examined in this study and demonstrate an observed association between packaging type and MP abundance. However, the findings do not establish packaging as the direct source of the detected particles, nor do they provide evidence of specific health effects. Further studies incorporating larger and geographically diverse sample sets, process-line monitoring, controlled packaging experiments, matrix-matched validation, and standardized analytical procedures are warranted to clarify the sources, occurrence, and potential significance of MP contamination in dairy products.

4.1. Practical implications

The findings of this study have practical implications for food-safety monitoring and dairy production. For policymakers and regulatory authorities, the detection of potential microplastic particles in commercially available milk highlights the need for continued monitoring and the development of standardized analytical approaches for microplastic assessment in dairy products. For dairy producers, strengthened quality-control practices throughout processing, packaging, storage, and handling may help minimize potential contamination. For consumers who regularly consume milk and dairy products, the findings indicate that these products may represent a potential dietary exposure pathway to microplastics. However, the present study does not establish direct human health effects, and the findings should therefore not be interpreted as evidence that milk consumption poses a demonstrated health risk. Further research and regulatory surveillance are needed to better characterize sources and potential exposure significance.

5. Conclusion

This study indicated the occurrence of potential microplastic (MP) particles in all 30 commercially available milk samples examined in Tehran, Iran, with concentrations ranging from 5 to 36 particles/L and a mean concentration of 18.4 ± 8.7 particles/L. Microscopic examination revealed a range of particle morphologies, including fragments, fibers, and spheres, while Micro-Raman spectroscopy confirmed the presence of several polymer types among the particles subjected to spectroscopic analysis, predominantly polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PET), polyamide (PA), and polyethersulfone (PES). The reported polymer composition therefore reflects the subset of particles successfully characterized by Micro-Raman spectroscopy and should not be interpreted as representative of all particles identified during microscopic screening.

MP abundance was significantly higher in plastic-bottled samples than in Tetra Pak cartons. However, this finding represents an association between packaging type and MP abundance and does not establish packaging as the direct source of the detected particles. Potential contributions from processing equipment, environmental contamination, handling, storage, and pre-processing sources cannot be excluded based on the present observational design. Determining the relative contribution of these sources would require process-line monitoring and controlled packaging-release experiments.

The analytical quality-control results showed polymer-specific recovery rates ranging from 82% to 96%, together with procedural and airborne blank measurements. These results provide information on analytical recovery and background contamination under the tested conditions. However, because the recovery experiments were conducted in ultrapure water rather than in milk, they should not be interpreted as complete matrix-matched validation of the analytical method. In addition, the detection threshold was derived from procedural blank measurements, and the analytical findings should be interpreted within the limitations of this quality-control approach.

The particle-size analysis showed that 72% of the measured particles were smaller than 100 μm, providing useful physicochemical information for future exposure and risk-assessment studies. Nevertheless, the present study did not evaluate biological uptake, tissue distribution, or toxicological effects. Therefore, the observed particle-size distribution does not provide direct evidence of human health effects.

Overall, this study provides evidence of potential MP occurrence in the commercial milk samples examined and demonstrates an observed association between packaging type and MP abundance. The findings should be interpreted within the limitations of the sample size, geographical coverage, source-tracking design, and matrix-specific analytical validation. Further research involving larger and geographically diverse sample sets, matrix-matched validation, process-line monitoring, controlled packaging experiments, and standardized analytical procedures is warranted to clarify the sources, occurrence, and potential significance of MP contamination in milk and other dairy products.

CRediT authorship contribution statement

Nazanin Jabbarzadeh: Sample collection, Laboratory analysis, Data curation, Investigation, Microscopy and spectroscopy analyses.

Mir-Hassan Moosavy: Supervision, Validation, Formal analysis, Writing - Review & Editing.

Marzie Hejazy: Conceptualization, Methodology, Project administration.

Asma Tarighi: Data interpretation, Visualization, Writing - Review & Editing, Writing - Original draft, Statistical analysis.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the University of Tabriz, Tabriz, Iran. Ethical approval was obtained before the study was conducted.

Consent for publication

Not applicable.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Funding

This research was financially supported by the Vice Chancellor for Research and Technology of the University of Tabriz. The funding body had no role in the study design, data collection, analysis, or interpretation of the results.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Handling Editor: Professor Aiqian Ye