Decadal accumulation of display materials in pristine soils of the Three Poles

Abstract

Identifying synthetic chemicals with long-range transport potential is central to chemical screening under the Stockholm Convention. The rapid global expansion of consumer electronics has spurred massive manufacturing of liquid crystal monomers and organic light-emitting materials. Although computational models predict persistent and semi-volatile behaviors for these compounds, field evidence documenting their multi-year presence in remote cryospheric sinks remains absent. Here we show that modern display materials have systematically infiltrated surface soils across Earth’s three polar regions—the Qinghai–Xizang Plateau, the Arctic, and the Antarctic. Across 55 soils collected between 2010 and 2022, we detected 42 target analogs at cumulative concentrations up to 3130 pg g⁻¹ dry weight (dw), reaching levels comparable to legacy persistent organic pollutants. Thirty-three analogs exhibited upward calculated annualized growth, yielding higher total concentrations in recent campaigns (26.8–3130 pg g⁻¹ dw) than earlier baselines (11.6–2090 pg g⁻¹ dw), predominantly driven by persistent fluorinated and cyano-substituted congeners. Spatial distributions reflect compound-specific air–soil partitioning may relate to volatility and soil organic carbon, while screening assessments reveal ecological risk quotients up to two orders of magnitude higher on the Qinghai–Xizang Plateau than in polar sites. These empirical findings indicate the environmental persistence and long-range transport of contemporary display chemicals, demonstrating a critical need to evaluate high-production optoelectronic materials under international regulatory frameworks.

Graphical abstract

Highlights

• •Display chemicals may widely infiltrate pristine soils across Earth’s Three Polar Regions.
• •Soil concentrations reach 3130 pg g⁻¹ dw, might matching legacy POPs and rising over a decade.
• •Distinct regional partitioning may relate to compound volatility and soil organic carbon.
• •Cryospheric accumulation urges evaluation of display materials under global POP treaties.

1. Introduction

Persistent organic pollutants (POPs) are regulated under the Stockholm Convention [1,2] due to their long-term adverse impacts on both ecosystems and human health. Assessing the environmental persistence and long-range transport potential (LRTP) of toxic and bioaccumulative substances is therefore essential for identifying and prioritizing potential POPs [2]. Attributed to their persistence and semi-volatility, these chemicals can transcend geographical boundaries and accumulate in pristine regions far from their emission sources under the global distillation [[3], [4], [5]]. Organochlorine pesticides (OCPs) and polychlorinated biphenyls (PCBs), for example, were detected in high-latitude polar regions in the 1960s and were considered to have reached these environments via atmospheric and oceanic transport [6,7]. Subsequent detection of candidate POPs in polar regions has provided evidence supporting their inclusion under the Stockholm Convention, and thirty-seven POPs have been listed as of 2026 [[8], [9], [10], [11], [12]]. Identifying new candidates nevertheless remains challenging, especially for high-production-volume chemicals whose environmental behavior is poorly understood. Because international regulatory frameworks, such as the Stockholm Convention, emphasize environmental persistence and LRTP as key criteria for evaluating chemical hazards, empirical evidence from remote regions is essential for assessing these substances [[13], [14], [15], [16]].

The three polar regions—the Arctic, the Antarctic, and the Qinghai–Xizang Plateau (QXP)—provide distinctive natural settings for assessing the environmental persistence and LRTP of emerging contaminants. Their limited direct anthropogenic inputs, extreme climates, and fragile ecosystems make them particularly sensitive indicators of contaminant transport and accumulation [17]. The QXP acts as a cold trap for semi-volatile compounds (PCBs and OCPs) due to its mountain cold condensation effect [18], whereas the Arctic and the Antarctic are recognized as global sinks for POPs owing to cold condensation at high latitudes [5]. Numerous legacy and emerging POPs have consequently been detected across the three polar regions, providing field evidence relevant to their persistence and transport. Yet data gaps persist for many high-production industrial chemicals, particularly those used in modern electronics, whose environmental fates remain insufficiently understood [[19], [20], [21]].

Liquid crystal monomers (LCMs) and organic light-emitting materials (OLEMs) are functional components of electronic display panels, and their production has surged over the past two decades as the electronics industry has expanded [22,23]. LCMs are usually classified as fluorinated biphenyls and their analogs (FBAs), cyanobiphenyls and their analogs (CBAs), and biphenyls/bicyclohexyls and their analogs (BAs). OLEMs are primarily polycyclic aromatic hydrocarbons bearing halogen atoms and are generally designed for thermal stability and optical performance. In silico studies indicate that both LCMs and OLEMs may exhibit toxic and persistent properties [19,24]. In addition, recent studies have identified LCMs and OLEMs in various environmental matrices (soil, water, and sediment) and in biota samples (zebrafish and mouse), and have also reported adverse effects of LCMs and OLEMs, including hepatotoxicity and endocrine disruption [[24], [25], [26], [27], [28], [29], [30]]. Their predicted semi-volatile properties and relatively long environmental half-lives (t_1/2,w ≥ 60 days) further suggest the potential for long-range transport [19,25,26,31]. However, field evidence from remote regions, especially the three polar regions, remains scarce due to the difficulty of sampling.

Here we investigated the occurrence of LCMs and OLEMs in soils from the three polar regions, using samples collected over a 5–10-year interval. We aimed to characterize their spatiotemporal distributions across remote regions, provide field-based evidence concerning their environmental persistence and LRTP, and evaluate their environmental fate and the ecological risks they pose to vulnerable ecosystems. As a field study to document the spatiotemporal variation of LCMs and OLEMs in the three polar regions, this study provides an initial assessment of their persistence and LRTP. The findings offer critical insights for chemical prioritization and risk management strategies in light of the global proliferation of electronics-related chemicals.

2. Materials and methods

2.1. Chemicals and sampling

We selected a total of 54 analogs—34 FBAs, 10 CBAs, 6 BAs, and 4 OLEMs—to represent the major structural classes of display-related chemicals. Selection was based on their reported industrial production and applications in display technologies, detection frequency in previous environmental studies [32,33], and structural diversity across physicochemical properties relevant to environmental fate [19,24,34]. Because quantitative production data were unavailable for many individual compounds, the analytes were not weighted by production volume. Hence, the selected analytes should be considered representative examples rather than an exhaustive set. Detailed physicochemical parameters are provided in Supplementary Table S1.

Standards for the 54 target compounds (purities > 98%) were purchased from Tokyo Chemical Industry Co., Ltd. (Shanghai, China), Bide Pharmatech Co., Ltd. (Shanghai, China), and Macklin Biochemical Co., Ltd. (Shanghai, China). Isotopically labeled PCB mixtures (P48-M-ES: ¹³C₁₂-PCB-28, -52, −101, −138, −153, and −180; P48-RS: ¹³C₁₂-PCB-70, -111, and −170) were obtained from Wellington Laboratories (Guelph, Ontario, Canada) and used as surrogate and internal standards. These standards were selected because their semi-volatile behavior and chromatographic properties are similar to those of LCMs. Moreover, they have been widely used as surrogate standards for hydrophobic organic contaminants in prior trace environmental analyses [33,35].

We collected 55 surface soil samples from the three polar regions during six Chinese scientific research expeditions. After vegetation, stones, and other surface debris had been removed, soil from the upper 0–5 cm was collected using pre-cleaned stainless-steel tools. On the QXP, samples were collected from Mount Shergyla (remote mountainous sites, S1–S9) and Lhasa (urban-influenced sites, S10–S14) to capture different anthropogenic gradients [36]. Arctic sites (S15–S18) were located near the Chinese Yellow River Station in Ny-Ålesund, Svalbard, and Antarctic sites (S19–S28) were located near the Chinese Great Wall Station. The geographical coordinates of all sampling sites are provided in Supplementary Text, section 1. Sampling campaigns were conducted in 2010 and 2021 on the QXP, in 2011 and 2016 in the Arctic, and in 2011 and 2022 in the Antarctic, thereby offering a broad temporal perspective on the presence of LCMs and OLEMs in remote environments.

Samples were wrapped in pre-cleaned aluminum foil, sealed in solvent-rinsed glass containers, transported frozen, and stored at −20 °C until analysis. Field blanks, consisting of pre-cleaned Na₂SO₄ exposed to sampling, transport, and storage, were included to evaluate potential contamination during sampling and handling. Frozen storage conditions were maintained throughout the sample preservation period to minimize the risk of contamination and analyte loss. No obvious abnormalities related to the storage duration were observed in the analytical results.

2.2. Pretreatment and instrumental analysis

Samples were prepared as described previously [33,35], with modifications. Lyophilized soil sample (5 g) was spiked with 1 ng of ¹³C₁₂-labeled PCBs (P48-M-ES) and extracted with dichloromethane/hexane (1:1, v/v) using an ASE 350 accelerated solvent extractor (Thermo Fisher Scientific, United States). Extraction was performed at 100 °C, 1500 psi over three cycles, with a static time of 5 min per cycle, a flush volume of 60%, and a nitrogen purge of 90 s. Extracts were concentrated and purified using a multilayer silica–alumina column packed, from bottom to top, with 1 g Na₂SO₄, 8 g silica gel, 1 g Na₂SO₄, 4 g Al₂O₃, and 1 g Na₂SO₄.

The extracts were further purified by gel permeation chromatography (GPC) to remove any high-molecular-weight interferences. The GPC fraction containing the target analytes was collected within a predefined elution window, which was established by analyzing a standard solution of all 54 target compounds. The elution range corresponding to these analytes was used to define the collection window, ensuring efficient removal of high-molecular-weight matrix components while retaining the target compounds. Recovery tests confirmed that the selected GPC window did not cause significant loss of target analytes. The final eluate was concentrated to 30 μL and supplemented with the internal standard (P48-RS) before instrumental analysis.

Target compounds were analyzed using a Trace 1300 gas chromatograph coupled with a TSQ 8000 Evo triple quadrupole mass spectrometer (Thermo Fisher Scientific, United States). Chromatographic separation was performed using a TG-5MS column (30 m × 0.25 mm internal diameter, 0.25 μm film thickness; Thermo Fisher Scientific, United States). The oven was held at 40 °C for 1 min, heated to 130 °C at 30 °C min⁻¹, to 250 °C at 5 °C min⁻¹, and to 300 °C at 10 °C min⁻¹, and then held at 300 °C for 10 min. The mass spectrometer was operated in electron ionization mode at 70 eV, with the ion source and transfer line maintained at 280 °C and 250 °C, respectively.

Data were acquired in selected reaction monitoring (SRM) mode with compound-specific transitions (Supplementary Table S2). Dwell times of approximately 10 ms were selected to obtain more than ten data points across each chromatographic peak. Compounds were identified by a retention time deviation of no more than ±0.1 min from the standards, the presence of at least two characteristic transitions, and transition ratios within ±20% of the calibration standards.

Soil total organic carbon (TOC) was measured using an SSM-5000A analyzer (Shimadzu, Japan) and was calculated as the difference between total carbon and inorganic carbon.

2.3. Quality assurance and quality control

A procedural blank was analyzed with every batch of eight samples, and field blanks were analyzed to assess contamination during sampling and storage. Chromatographic performance was evaluated to minimize co-elution, and no major interferences that would affect quantification were observed. Representative samples were re-analyzed to verify the chromatographic separation, peak integration, and ion-transition consistency. Representative qualitative and quantitative chromatograms of selected analytes are provided in Supplementary Fig. S1.

For analogs detected in the blanks, method detection limits (MDLs) were defined as the mean blank concentration plus three times its standard deviation (SD); for analogs undetectable in the blanks, MDLs were calculated as three times the SD of the lowest standard level (5 pg μL⁻¹) [37]. MDL values ranged from 1.5 to 7.2 pg g⁻¹ for all targets in this study (Supplementary Table S3). For analytes identified in both blanks and actual samples, detections were considered valid only when concentrations exceeded the corresponding MDL. Concentrations below the MDLs were treated as non-detects (ND).

Six solvent-based calibration curves were prepared over a concentration range of 5–500 pg μL⁻¹. Method recovery was evaluated by fortifying target-analyte-free soil at 0.2, 2, and 18 ng per 5 g of soil (n = 10 per group). Recoveries across the 54 compounds ranged from 80.1% to 140% (Supplementary Tables S4 and S5), broadly consistent with values previously reported for soils from other regions (70–148%) [33,35]. Values exceeding 120% suggested potential matrix enhancement or calibration bias and may have led to the overestimation of some analytes.

Meanwhile, the recoveries of the internal standard (P48-M-ES: ¹³C₁₂-28, ¹³C₁₂-52, ¹³C₁₂-101, ¹³C₁₂-138, ¹³C₁₂-153, and ¹³C₁₂-180) ranged from 67.3% ± 2.5% to 88.0% ± 3.8%, which indicated the high reliability of the entire analytical procedure (Supplementary Table S6). In the real sample analysis, the recoveries of the internal standard (P48-M-ES) ranged from 59.2% ± 8.0% to 63.5% ± 9.9% (Supplementary Table S6), indicating moderate variability in analytical performance. PCB-based surrogates were used to correct for procedural losses and exhibited acceptable performance across the target analytes despite structural diversity.

2.4. Statistical analysis

Statistical analyses were performed using RStudio (version 2024.12.1 + 563, Posit Software, United States). Data normality and homogeneity of variance were evaluated before one-way analysis of variance (ANOVA). Differences among the groups were analyzed using one-way ANOVA, and correlations between the variables were assessed using Spearman’s rank correlation analysis. Statistical significance was defined as p < 0.05.

3. Results and discussion

3.1. Occurrence of LCMs and OLEMs in the three polar regions

We detected 42 of the 54 target analogs—25 FBAs, 9 CBAs, 4 BAs, and 4 OLEMs—in soils from the three polar regions, of which 21 had detection frequencies greater than 30% (Supplementary Fig. S2). The sum concentrations of the 42 detected compounds ranged from 61.8 to 2860 pg g⁻¹ dry weight (dw) in the QXP, with a mean of 911 pg g⁻¹ and a median of 656 pg g⁻¹. Corresponding concentrations were 24.7–1700 pg g⁻¹ dw in the Arctic (mean, 440 pg g⁻¹; median, 175 pg g⁻¹) and 11.6–3130 pg g⁻¹ in the Antarctic (mean, 423 pg g⁻¹; median, 140 pg g⁻¹) (Fig. 1). These findings indicated the widespread occurrence of display-related contaminants in remote polar and high-altitude soils.

Fig. 1.

Concentrations of LCMs and OLEMs in soil samples collected from three polar regions during different sampling campaigns. a–c, Concentrations of individual LCMs and OLEMs in soil samples from the Qinghai–Xizang Plateau (QXP; Lhasa and Mount Shergyla) in 2010 and 2021 (a), the Arctic in 2011 and 2016 (b), and Antarctica in 2011 and 2022 (c). S1–S28 are sampling-site identifiers: S1–S9 represent remote mountainous sites on Mount Shergyla; S10–S14 represent urban-influenced sites in Lhasa; S15–S18 represent sites near the Chinese Yellow River Station in Ny-Ålesund, Svalbard; and S19–S28 represent sites near the Chinese Great Wall Station in Antarctica. Detailed coordinates for all sampling sites are provided in the Supplementary Text. Sites absent from a given campaign were not sampled because of field-access difficulties. Repeated site identifiers indicate separate samples collected in the vicinity of the same site during the same campaign. d, Comparison of the summed concentrations of 42 LCMs and OLEMs (C_{Σ42LCMs + OLEMs}) among regions and sampling campaigns. Colors denote different functional groups of analogs. Bars, error bars, and points represent the mean, standard deviation, and individual values of concentrations. LCM, liquid crystal monomers; OLEMs, organic light-emitting materials; FBAs, fluorinated biphenyls and their analogs; CBAs, cyanobiphenyls and their analogs; BAs, biphenyls/bicyclohexyls and their analogs. Sample labels follow the format region_sampling year; for example, QXP_2010 refers collectively to all samples collected from QXP in 2010.

This study developed a field dataset concerning the presence of LCMs and OLEMs in soils from remote polar and high-altitude regions. In general, the concentrations measured in both the Arctic and the Antarctic were substantially lower than those previously measured in source-related areas such as e-waste recycling areas (Qingyuan, Guangdong), regions home to display manufacturers (Guangzhou, Guangdong), and urbanized environments in China (Supplementary Table S7) [33,35]. However, the concentrations obtained in some QXP samples were comparable to those reported in those more impacted environments [33,35]. The comparable mean concentrations between the Arctic and the Antarctic in this study should also be interpreted in light of the different sampling years (2016 for the Arctic and 2022 for the Antarctic). Given the ongoing release of these “chemicals in use,” it cannot be ruled out that the Arctic may have received higher inputs of LCMs and/or OLEMs by 2021, with implications for remote ecosystems.

Based on cross-study comparisons, the measured concentrations were comparable to previously reported levels of legacy and emerging POPs in remote regions (Supplementary Table S8 and Fig. S3) [38]. The C_{∑42LCMs + OLEMs} in this study (61.8–2860 pg g⁻¹ dw in the QXP and 11.6–3130 pg g⁻¹ dw in the other two polar regions) was comparable with those reported for several classes of POPs, including PCBs, PBDEs, and NBFRs, in those comparable remote environments [4,36,[39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57]]. Despite the known differences in sampling period, depth, analyte coverage, methodology, and metrics among the considered studies, these comparisons provide a useful benchmark for evaluating the environmental relevance of emerging LCMs and OLEMs relative to legacy POPs. Previous studies have suggested that several fluorinated LCM analogs possess semi-volatile properties, bioaccumulation potential, and toxicological effects, including dioxin-like activity [23,24,26,52]. The detection of these compounds in remote soils therefore aligns with growing concerns about their environmental persistence and potential for long-range transport. Measurements across additional environmental media, combined with mechanistic transport studies, are required to establish their global environmental fate and ecological implications.

3.2. Composition of LCMs and OLEMs in soils

Class-level profiles differed among the three regions. Summed concentrations followed the order ∑FBAs > ∑CBAs > ∑BAs > ∑OLEMs in the QXP, ∑CBAs ≥ ∑FBAs > ∑BAs > ∑OLEMs in the Arctic, and ∑FBAs > ∑BAs > ∑CBAs > ∑OLEMs in the Antarctic (Fig. 2a). FBAs dominated most samples, whereas OLEMs contributed the lowest proportions. This pattern may reflect, at least in part, the longer and larger-scale industrial production and application of FBAs following the rapid expansion of liquid crystal display technologies after 2000, whereas the commercial use of OLEMs only became widespread more recently (around 2010) [22,23]. However, only four OLEMs were included in the analytical panel, and their lower contribution should not be interpreted as evidence of lower environmental abundance across the wider chemical class.

Fig. 2.

Composition profiles of LCMs and OLEMs in soils from the Qinghai–Xizang Plateau, the Arctic, and Antarctica. a, Relative contributions of four functional groups to the total concentrations of LCMs and OLEMs in each region and sampling campaign. Bars and error bars represent mean and standard deviation values. b, Heatmap showing the normalized relative abundances of individual analogs across soil samples. Rows represent individual analogs, and columns represent samples clustered using Euclidean distance and complete linkage. The dendrogram scale indicates Euclidean distance, calculated from normalized values rescaled to the range −1 to 1. Colors indicate normalized relative abundance according to the adjacent scale, and colored side bars identify the four functional groups. Branch colors in the column dendrogram distinguish three clusters obtained by cutting the dendrogram into three groups; no predefined sample-group information was used for clustering. S1–S28 are sampling-site identifiers: S1–S9 represent remote mountainous sites on Mount Shergyla; S10–S14 represent urban-influenced sites in Lhasa; S15–S18 represent sites near the Chinese Yellow River Station in Ny-Ålesund, Svalbard; and S19–S28 represent sites near the Chinese Great Wall Station in Antarctica. Detailed coordinates for all sampling sites are provided in the Supplementary Text. Sample labels follow the format region_sampling year_site identifier; for example, QXP_2021_S4 denotes a sample collected at site S4 on the Qinghai–Xizang Plateau in 2021. Labels without a site identifier, such as QXP_2021, refer collectively to all samples from that region and sampling campaign. QXP, Qinghai–Xizang Plateau. LCM, liquid crystal monomers; OLEMs, organic light-emitting materials; FBAs, fluorinated biphenyls and their analogs; CBAs, cyanobiphenyls and their analogs; BAs, biphenyls/bicyclohexyls and their analogs.

The relatively higher contribution of BAs in the Antarctic may reflect differences in the physicochemical properties and regional environmental conditions. Some BA analogs exhibit physicochemical characteristics favorable for atmospheric transport [25]. While the present dataset does not fully elucidate the underlying transport mechanisms, the observed compositional differences are consistent with variability in source profiles, meteorological conditions, deposition processes, and sampling years among the assessed regions [58,59]. Compositional differences between the windward and leeward sides of Mount Shergyla further suggest that regional meteorological conditions may have influenced the environmental distribution of the investigated compounds [36]. In addition, the TOC content in the soils showed positive correlations with the concentrations of analogs with logarithm of the octanol–water partition coefficient (logK_ow) > 5 (Supplementary Fig. S4, analogs n = 39, p < 0.05), suggesting that soil organic matter may contribute to the retention of relatively hydrophobic compounds in soils.

Different patterns were also observed among the individual analogs across the regions over the sampling periods (Fig. 2b–Supplementary Table S9–S10). The FBA analog 3bcHdFB, a commonly reported LCM component of display panels, dominated QXP samples, accounting for 63.6–82.4% of the measured concentrations. In the Arctic, the CBA analog 2OCB accounted for 29.0% of the concentrations in 2011, whereas the BA analog MePVbcH accounted for 31.8% in 2016. MePVbcH was also a major contributor in the Antarctic soils, accounting for 20.8–42.0%. Concentrations of several dominant analogs in remote regions were comparable to those previously reported in more impacted environments (Supplementary Table S7) [33,35]. For example, the mean concentration of 3bcHdFB in QXP soils exceeded the value reported for agricultural soils from an industrialized region of China (69 pg g⁻¹ dw) [35]. Similarly, the concentrations of MePVbcH detected in certain recent polar samples approached the levels reported in e-waste recycling and display manufacturer areas (320 pg g⁻¹ dw) [33].

The dominance of compounds such as 3bcHdFB, 2OCB, and MePVbcH likely reflects the combined influence of the production volume, usage pattern, environmental release, and physicochemical behavior. For instance, MePVbcH has physicochemical properties (i.e., relatively high logarithm of the octanol–air partition coefficient (logK_oa) and logK_ow) that may favor partitioning into organic matter while still allowing atmospheric transport under certain environmental conditions (Grasshopper effect) [26,60]. The temporal differences in the analog composition among the regions may also reflect changes in the industrial usage patterns of display-related chemicals over the past decade, as well as differences in the transport and environmental partitioning behavior among the compounds [22,26].

3.3. Spatiotemporal distribution

Synthetic organic contaminants detected in remote environments may originate from anthropogenic emissions and long-range environmental transport. Concentrations of LCMs and OLEMs did not differ significantly between soils collected near the Arctic and Antarctic research stations (S16 and S23) and those collected at more distant sites, or between Lhasa and Mount Shergyla (one-way ANOVA, p > 0.05). These observations suggested that local sources were not the dominant factor controlling the regional distribution patterns observed in this study. Indeed, while local contributions from research stations cannot be entirely ruled out, the overall dataset supports the LRTP as a primary pathway for these contaminants to reach remote regions.

Prior studies have predicted that Arctic contamination and bioaccumulation potential (ACBAP) for several LCM analogs based on their physicochemical properties [26,34]. In addition, many LCMs have been reported to exhibit semi-volatile behavior and the ability to partition into the atmosphere [25]. In contrast, the atmospheric fate and transport of OLEMs remains limited. Their potential semi-volatility is primarily inferred from their structural similarities with other aromatic organic compounds and therefore requires further experimental confirmation. Yet the detection of both LCMs and OLEMs in remote soils from the QXP, Arctic, and Antarctic is nevertheless consistent with the hypothesis that some display-related chemicals can undergo regional or long-range environmental transport.

At Mount Shergyla, concentrations of several analogs exhibited altitude-related variation and were positively correlated with TOC content along the 2600–4500 m elevation gradient (Supplementary Figs. S5 and S6). These relations suggested that soil organic matter may influence the retention of hydrophobic analogs in mountain soils. Moreover, the different analog compositions between the windward and leeward sides of Mount Shergyla may also reflect the influence of regional meteorological conditions and atmospheric circulation patterns [[61], [62], [63]]. The Indian monsoon can affect the transport and deposition of POPs in the QXP [62,63], potentially carrying contaminants from South Asian source regions to the southern plateau. On the leeward side, the local mountain–valley wind circulation may further influence the deposition patterns. Direct atmospheric measurements or trajectory analyses were beyond the scope of the present study, although the observed spatial patterns were consistent with these transport-related hypotheses.

Compound profiles also differed among the three regions (Fig. 2b), suggesting that regional environmental conditions and transport pathways may influence the distribution of individual analogs. In the Arctic, for instance, atmospheric transport from mid-latitude source regions has previously been identified as an important pathway for semi-volatile contaminants [[64], [65], [66]]. In Antarctica, oceanic and atmospheric circulation systems may contribute to contaminant delivery despite the lack of local sources [67].

Temporal differences were also observed between the earlier and more recent sampling campaigns (Fig. 1). Summed concentrations of the 42 detected compounds were generally higher in samples collected during 2016–2022 than in those collected during 2010–2011. Mean MePVbcH concentrations reached 203 pg g⁻¹ dw in Arctic soils collected in 2016 and 279 pg g⁻¹ in Antarctic soils collected in 2022, exceeding those measured during the earlier sampling periods. However, 3bcHdFB remained the dominant analog in the QXP during both sampling periods. These observations may reflect changes in the production, use, disposal, and/or environmental release of display-related chemicals over the past decade.

Previous modeling research identified four physicochemical criteria associated with relatively high ACBAP: logK_oa + logarithm of the air–water partition coefficient (logK_aw) ≥ 3.5; logK_oa ≥ 6, 0.5 ≥ logK_aw ≥ −7; and logK_aw ≤ −1.78 × logK_oa + 14.56 [26,31]. Ten target analogs detected in the three polar regions met these criteria. To further explore possible accumulation behavior, we compared changes in the apparent concentrations of individual analogs with the reported global market growth rate for display-related materials (approximately 16%) [25,68]. Due to the current lack of detailed emission inventories for individual analogs, this comparison serves as only an exploratory framework for evaluating potential accumulation tendencies. Analogs that showed concentration increases exceeding the reported market growth rate were tentatively considered to exhibit relatively stronger apparent accumulation tendencies in remote regions, with the understanding that environmental release, transport efficiency, degradation, and usage patterns vary among compounds and that the growth rate was derived from only two sampling periods. Among the detectable analogs, many compounds with relatively greater apparent accumulation tendencies possessed physicochemical properties within the ranges of 8 < logK_oa < 12, 5 < logK_ow < 10, and 300 < molecular weight < 400 g mol⁻¹.

Exploratory correlation analyses showed a weak negative association between concentration changes and logK_ow on the QXP (R = −0.3, p = 0.1) and weak positive associations between concentration changes and logK_aw in the Arctic (R = 0.4, p = 0.1) and Antarctic (R = 0.5, p < 0.05) (Fig. 3). While two of these correlations did not achieve statistical significance, the observed patterns align with a conceptual framework where the soil retention processes associated with hydrophobicity may exert a relatively greater influence in the QXP, whereas the properties associated with atmospheric transport may play a more prominent role in the Arctic and Antarctic. Given the exploratory nature of this analysis and its limited temporal resolution, these patterns should be considered indicative rather than definitive. Future studies incorporating atmospheric monitoring, multimedia partitioning, and source apportionment analyses are required to verify these hypotheses.

Fig. 3.

Relationships between temporal increases in LCM and OLEM concentrations and physicochemical properties. Relationships between the calculated growth rates of individual analogs and the logarithm of the octanol–water partition coefficient (logK_ow) on the Qinghai–Xizang Plateau (left) or the logarithm of the air–water partition coefficient (logK_aw) in the Arctic (middle) and Antarctica (right). Growth rate was calculated as the annualized percentage change in measured concentration between the earlier and later sampling campaigns; analogs not detected during the earlier campaign were excluded. Colors identify the four functional groups. Solid lines show fitted relationships and shaded regions represent 95% confidence intervals; n, number of analogs included compounds with positive annualized growth rates; R, Pearson’s correlation coefficient; LCM, liquid crystal monomers; OLEMs, organic light-emitting materials; FBAs, fluorinated biphenyls and their analogs; CBAs, cyanobiphenyls and their analogs; BAs, biphenyls/bicyclohexyls and their analogs.

3.4. Ecological risk in the three polar regions

Cryospheric ecosystems in the three polar regions are environmentally sensitive and potentially vulnerable to POP accumulation [69]. Given the ongoing use and release of display-related chemicals, the occurrence and potential accumulation of LCMs and OLEMs in remote regions warrant preliminary ecological evaluation. Prior studies have reported cytotoxicity and predicted toxicological effects in several LCM analogs, particularly FBAs and CBAs, suggesting that the potential environmental impacts of these analogs should not be ignored [32,70]. We therefore conducted a screening-level ecological risk assessment using risk quotient (RQ) values based on the measured concentrations and predicted no-effect concentrations (PNECs) (Fig. 4) [4,71,72]. The detailed methodology for the PNEC derivation is provided in the Supplementary Text, including the ecological structural activity relationship (ECOSAR) parameters, species and endpoints, assessment factor justification, and complete calculation results. The calculated RQ values for the QXP were generally one to two orders of magnitude higher than those for the Arctic and Antarctic, indicating the relatively greater potential ecological concern in the QXP. We derived PNEC values from aquatic toxicity data predicted by ECOSAR using the equilibrium partitioning method and applied standard assessment factors to establish the screening-level PNECs (Supplementary Text, section 2). As soil-specific toxicity data for LCMs and OLEMs are currently unavailable, the present RQ analysis should be regarded as only a preliminary prioritization exercise (Supplementary Text, section 3). In addition, the physicochemical properties of many LCM analogs, including high hydrophobicity and low solubility, may have introduced further uncertainty into the toxicity predictions and environmental exposure estimates.

Fig. 4.

Risk quotients of selected LCMs and OLEMs in soils from the Qinghai–Xizang Plateau, the Arctic, and Antarctica. Calculated risk quotients (RQs) for analogs (RQ > 10⁻¹²) in soil samples collected during the different sampling campaigns. Bars show mean RQs, and error bars indicate standard deviations across soil samples collected within each region and year. Colors identify the four functional groups. Ecological risk was classified as high (RQ ≥ 1), medium (0.1 ≤ RQ < 1), and low (RQ < 0.1). Within each panel, analogs are arranged in decreasing order of mean RQ. Sample labels follow the format region_sampling year; for example, QXP_2010 refers collectively to all samples collected from Qinghai–Xizang Plateau in 2010. LCM, liquid crystal monomers; OLEMs, organic light-emitting materials; FBAs, fluorinated biphenyls and their analogs; CBAs, cyanobiphenyls and their analogs; BAs, biphenyls/bicyclohexyls and their analogs.

Several compounds in the QXP exhibited increasing occurrence and elevated relative RQ values when compared with those in the Arctic and Antarctic. While most RQ values remained below 0.1 and were generally lower than those previously reported for novel brominated flame retardants (1.96 × 10⁻⁸ to 0.062) [4], the observed temporal differences and increasing detection frequencies indicated that the continued environmental release of these compounds may gradually increase ecological exposure in remote regions. On that basis, future studies should focus on the acquisition of experimentally measured soil toxicity data, improved characterization of environmental exposure, and multimedia fate assessment for LCMs and OLEMs to better constrain their environmental risk.

We further examined the potential toxicological implications of LCMs based on previously reported predictive studies [32,70]. Wu et al. developed high-throughput screening approaches that combined classification models and quantitative structure–activity relationship models to evaluate the potential hepatotoxic effects of LCMs [70]. Their results suggested that LCM analogs containing fluorine or cyano functional groups may exhibit relatively higher cytotoxicity potential, attributed to their strong binding affinity with antioxidant enzymes and high lipophilicity [70]. Xu et al. evaluated the potential soil health effect of LCMs by means of molecular docking and molecular dynamics simulation methods from multiple mechanistic dimensions [32]. Analogs containing fluorine or cyano groups were again predicted to exhibit relatively stronger interactions with soil-related biological targets. In this study, twenty-six analogs (excluding OLEM-21) that exhibited predicted persistence and LRTP were also screened using these previously published predictive frameworks. Among them, eight analogs (i.e., 2OdF3B, 3cHtFB, 3bcHtFB, 3cHdFB, 2F3T, 3OdFP3bcH, tFPO-CF2-tF3T, and 5CT), mainly containing fluorine or cyano substituents and multiple aromatic rings, were predicted to exhibit relatively higher cytotoxicity and potential soil health effects. While these predictions were derived from computational models, they nevertheless provide a strong basis for prioritizing these analogs for further toxicological study. Indeed, their persistence and semi-volatility are consistent with POP-like properties, suggesting that certain LCMs may behave similarly to persistent organic contaminants.

4. Conclusions

This study generated field datasets describing the presence of LCMs and OLEMs in soils from the three polar regions. Their detection in remote polar and high-altitude environments demonstrates their potential to be transported beyond primary emission regions. The temporal differences between the earlier and more recent sampling campaigns in this study may indicate increasing environmental occurrence associated with the expanding global use of display technologies, although direct interpretation of this trend is constrained by the differences in sampling years and limited temporal coverage. Several dominant analogs exhibited physicochemical properties previously associated with both environmental persistence and transport potential, particularly the fluorinated and cyano-substituted compounds. The preliminary ecological risk screening further suggests that some analogs may warrant additional attention in sensitive cryospheric ecosystems, especially given the current lack of soil-specific toxicity data. Overall, the results of this study provide an initial basis for future investigations into the environmental behaviors, transport pathways, and ecological impacts of LCMs and OLEMs. Studies integrating atmospheric measurements, multimedia modeling, emission inventories, and ecotoxicological assessments are required to more thoroughly evaluate the environmental significance of these emerging contaminants.

CRediT authorship contribution statement

Yanfen Hao: Writing – original draft, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. Jing Xu: Visualization, Investigation, Data curation. Tiantian Han: Investigation, Data curation. Yizheng Ge: Investigation, Data curation. Jiachen Yan: Data curation. Thanh Wang: Writing – review & editing. Pu Wang: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Qinghua Zhang: Supervision, Conceptualization. Duo Bu: Supervision, Conceptualization. Yong Liang: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization.

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.

Appendix A. Supplementary data

The following is the Supplementary data to this article: