Occurrence of tire-derived microplastics (TMPs) focusing on driving behavior

Microplastics (MPs) are defined as plastic particles with a diameter of 5 mm or smaller, and their release into aquatic environments has been reported (Andrady 2011). MPs can be classified into two major categories: primary and secondary MPs. Primary MPs are particles that are already 5 mm or smaller at the time of manufacture, which include microbeads found in facial cleansers and cosmetics. Secondary MPs are used plastic products that have been improperly disposed of, and the diameter has been reduced to 5 mm or smaller after crushing and fragmenting via ultraviolet light, heat, wind, etc. (Takada 2018). Plastic product debris is considered the main source of MPs along the coast of Japan (Isobe 2016). Future estimates suggest that by 2060, the density of MPs in certain areas of the Pacific Ocean could reach four times the current level, which could adversely affect ecosystems (Isobe et al. 2019). Ecological impacts have been reported (Ma et al. 2016; Zhang et al. 2017), and it has been confirmed that marine plastics contain various toxic chemicals derived from antioxidants and flame-retardant additives (Teuten et al. 2009), which remain in MPs. Thus, marine runoff containing MPs is considered a problem worldwide. Among MPs, tire-derived MPs (TMPs) are estimated to account for 60% of the total secondary MPs emitted into the aquatic environment (Lassen et al. 2015), making the emissions that stem from terrestrial areas a problem as well. Furthermore, it is estimated that approximately 34% of the emitted coarse tire wear particles and 30% of the emitted coarse brake wear particles were deposited in the World Ocean (Evangeliou et al. 2020). Reports from the Netherlands estimate that 1,800 tons of tire wear MPs are released into surface water annually (National Institute of Public Health & the Environment Netherland 2016). Moreover, it has been noted that the toxicity of tire wear particles may be different from that of tire rubber itself (Wik & Dave 2009); therefore, it is important to investigate both the quantity and quality of TMPs separately from various other MPs.

There are reports (Carr et al. 2016; Tanaka et al. 2019) that more than 99% of MPs can be removed by sewage treatment plants. However, the removal rate cannot reach 100%, and MPs are also detected in treated sewage effluent. Furthermore, MPs are also transported to the ocean via overflow during rainy weather conditions (Takada 2018). In addition, newly planned sewage systems are based on the principle of a separate sewer system, which enables MPs to pass through sewage systems without undergoing any of the treatment that household wastewater receives. Therefore, it is necessary to understand the occurrence of TMPs from roads as a source of pollution.

There are a few studies that focus on tire dust generation on roads (Järlskog et al. 2020a, 2020b; Klockner et al. 2021; Kovochich et al. 2021). Those studies confirm traffic as an important source of MPs (Järlskog et al. 2020b), reveal different size distributions among sampling sources (Järlskog et al. 2020a; Klockner et al. 2021; Kovochich et al. 2021), and propose a model to predict emissions (Lundberg et al. 2021). Various methods have been employed for the identification of particles of tire origin. Some of those studies rely on microscopic visual inspection, followed by instant morphological analysis, such as burning (Järlskog et al. 2020a), using a modeling approach to estimate the source (Karanasiou et al. 2014), measuring marker substances such as zinc (Klockner et al. 2021), or visual inspection with identification by GCMS (Jung & Choi 2022). However, conventional techniques relying on morphological analysis have limitations in terms of estimating the actual occurrence of TMPs because tires were originally a mixture of complex materials (Baensch-Baltruschat et al. 2020; Rausch et al. 2022). Moreover, tires are expected to absorb organic materials in the environment (Huffer et al. 2019). Therefore, the emission of TMPs should be investigated using a suitable technique.

A micro Fourier transform infrared spectroscopy (μFTIR) is used to measure MPs because the technique provides useful information about the materials by referencing the spectrum database (Tagg et al. 2015). Recent studies have used FTIR to measure MPs on road (Okamoto et al. 2019; Roychand & Pramanik 2020; Yukioka et al. 2020) or in water samples (Leads & Weinstein 2019; Ziajahromi et al. 2020) and have successfully detected rubber or tires in those samples. Another study used a modeling approach to estimate the occurrence of vehicle tires (Prenner et al. 2021). However, those studies have only reported the abundance of MPs and did not conduct further analysis, such as particle size distribution, composition or spatiotemporal distributions. Considering the generation of tire particles, vehicle driving behavior such as braking is expected to have a significant impact. However, to the best of our knowledge, there are few cases in which the temporal and spatial distributions of these materials on actual roadways have been examined. Moreover, no research has investigated the impact on MP generation from tires from a traffic engineering perspective.

With this background, this study surveyed the occurrence of TMPs on roads by μFTIR and focused on driving behavior. More specifically, samples were taken at different distances from the stop line to compare the number of TMPs. In addition, vehicle driving behavior was recorded on video and analyzed to calculate acceleration and deceleration. Finally, relations between the number of TMPs and vehicle behaviors were discussed.

Road dust was sampled by using a vacuum cleaner (CL102DW Makita) following previous studies on MPs (Yukioka et al. 2020). Briefly, road dust was collected into a disposable filter from a 6-m² square road area in a driving lane (2 m length by 3 m width) during 1 min of vacuuming. The recovery and size distribution of tire particles by this operation is further explained in the supporting material. Sampling was conducted when vehicle traffic was interrupted. The 6-m² square was set several distances from a stop line. The nozzle, tube and filter of the vacuum cleaner were changed for each sample to avoid contamination. With these procedures, road dust was collected in a disposable filter, transported to the laboratory in a sealed container, and provided for further analysis after passing through a 4.75-mm stainless steel sieve (JIS Z 8801 test sieves).

Sampling sites were determined based on traffic volume survey results from the road traffic census (Ministry of Land, Infrastructure, Transport and Tourism 2015). Details on the sampling sites are provided in Table 1. In this study, sampling was conducted in (i) Minami-Osawa, Hachioji city, (ii) Kotta, Tama city and (iii) Mahikizawa, Tama city. All sampling locations have two lanes traveling in one direction. Sampling was conducted at a specific distance from the stop line in both the right and left driving lanes to find a relationship with driving behavior. Since Japan is a left-hand traffic country, the left lane is basically the driving lane and the right lane is the passing lane when there are two lanes on one side of the road. Sampling in Kotta was conducted on 3 different days. To check the impact of various parameters, such as the number of antecedent dry days, some samples were repeatedly taken from the same spots on a different day. Specifically, a sample at 150 m spot of sampling location 4 (i.e., sampling No. 4) and a sample at 150 m spot of sampling location 2 (i.e., sampling No. 2) were collected from exactly the same spot on a different day. Similarly, samples from sampling locations 3 and 4 were collected from exactly the same spots at 30, 50 and 70 m in the left lane on different days.

Based on the particle size, most TMPs were 105 μm or larger. There was variability in the number of TMPs above that size, which could be a result of traffic volume or driving behavior. In contrast, there was less variability in the number of TMPs that were 105 μm or smaller, which will be discussed in a later section. Briefly, those numbers were within the range of 9–16, except for three samples, which were from 30 and 50 m in the right lane of sampling No. 4 and 30 m in the right lane of sampling No. 5.

The range of the number of MP was similar to or larger than that from previous studies using μFTIR. A study by Yukioka et al. (2020) reported an abundance of rubber MPs in the range of 0.4–4.5 pieces/m² in road dust from Japan, Vietnam and Nepal. In our study, the range of TMPs/m² is approximately 2–9 pieces/m² for the whole size range and 0.2–7.7 pieces/m² for the 105 μm or larger range. The increased number of TMPs in this study may be due to improvement of the detection methodology of this study, which used the actual tire spectrum as a reference. Leads & Weinstein (2019) used tire as a reference spectrum and reported the number of TMPs in a 0.25-m square of intertidal sediment as 198, 94 and 19 for size ranges of 63–149, 150–499 and ≥500 μm, respectively. Although the numbers by Leads & Weinstein (2019) cannot be directly compared, the use of tires as a reference spectrum would improve the probability of identification using μFTIR. One of the studies (Järlskog et al. 2020b) using investigation techniques other than μFTIR reported a considerably larger number of TMPs (on the order of hundreds/m²) that were 100 μm or larger in road dust in Sweden. Although it is also difficult to directly compare due to differences in road characteristics in different countries, the higher number of TMPs in their studies may be due to differences in investigation techniques, which might identify a wider range of particles as TMPs. In fact, the ratio of TMPs and other MPs in the study (Järlskog et al. 2020b) is approximately 100:1, which is approximately 1:1 or at least the same magnitude in other studies (Leads & Weinstein 2019; Yukioka et al. 2020). Further investigation will be needed to estimate the actual number of TMPs from road dust.

Similarly, comparing the results at sampling No. 4, Kotta, to those at sampling No. 5, Mahikizawa, sampling No. 5 experienced 1 more day of clear weather than sampling No. 4 and less than half the traffic volume. The total number of TMPs was 250 from sampling No. 4 in Kotta and 183 in sampling No. 5 in Mahikizawa. This result suggests that although the number of antecedent dry days was smaller at sampling No. 4, the larger traffic volume in sampling No. 4 resulted in a larger number of TMPs. These results indicate that the traffic volume is one of the major factors impacting the number of TMPs.

Similarly, we compared the number of TMPs between three spots in the left lane of the No. 3 sampling site and at 30, 50 and 70 m for the No. 4 sampling site. The number of antecedent dry days at the No. 3 sampling site was 8 days, and the number of antecedent dry days at the No. 4 sampling site was 4 days. Despite the difference in the number of antecedent dry days, the number of TMPs derived from tire dust was similar. This result suggests that the number of TMPs can reach a plateau approximately 4 days after the most recently observed rainfall event.

In fact, a previous study (Amato et al. 2012) summarized the temporal variability in road dust 10 μm or smaller in size and demonstrated that the number of TMPs exponentially increased over 3 days after rainfall ceased and remained the same thereafter. Therefore, it was determined that the amount of TMPs does not change after a certain period has passed from the most recent rainfall event, even for tire dust-derived TMPs.

Analysis of the relationship with driving behavior was conducted mainly with reference to knowledge of dilemma zones in traffic engineering. The dilemma zone refers to a zone where the relationship between the speed at that time that the traffic light turns yellow and the distance to the stop line creates hesitation as to whether to continue or stop. It should be mentioned that the Japanese traffic signal system uses the all-red system, which gives drivers a choice regarding yellow signals.

First, we examine the results in the left lane for sampling No. 3 and No. 4. Comparing the number of TMPs at 70 m and other distances in the left lane of sampling No. 3 and No. 4, the number of TMPs at 70 m is approximately 28% higher than the average of TMPs at other distances. Since the legal speed limit on Metropolitan Route 158 is 60 km/h, it was assumed that the vehicles were traveling at 60 km/h. In this case, it could be considered that the driver passes the stop line at the start of a yellow light if the distance to the stop line is within 0–50 m, but if the distance exceeds 60 m, then the driver would stop without hesitation. Therefore, in the present case, it was assumed that stopping is the main behavior at 70 m, while the majority of vehicles pass at 50 and 30 m.

The results for the right lane were similar to those for the left lane, and the highest number of TMPs was found at 70 m, approximately 83% higher than that at other distances on average. There was a minor difference in the trend from the left lane. The right lane exhibited the fewest TMPs at 30 m, while the left lane exhibited the fewest TMPs at 50 m. One probable reason for this finding may be the difference in vehicle behavior between the left and right lanes at the sampling location. The number of TMPs was also higher in the right lane than in the left lane, suggesting that the difference in traffic volume also yielded an effect.

Next, we examined the results for sampling No. 5. Since the legal speed limit on Metropolitan Route 18 is 50 km/h, it was assumed that the vehicles were traveling at 50 km/h. In this case, it was expected that stopping was the main behavior at 50 and 70 m and that the majority of vehicles passed at 30 m. As a result, the highest number of TMPs was observed in the left lane at 50 m, indicating the order of 50 m > 70 m > 30 m. Comparison of Route 158 and Route 18 suggests that the locations where TMPs are greatest differ depending on the legal speed limit. On Route 158 at a legal speed of 60 km/h, TMPs were most frequently found at 70 m from the stop line, while on Route 18 (at a legal speed of 50 km/h) TMPs were found at 50 m. This result suggests that the location where TMPs occur most frequently is influenced by the legal speed of the road.

Figure S1 shows the number of TMPs at 105 μm or smaller. As mentioned in the overall results, fewer TMPs were observed in the right lane at 30 and 50 m at sampling No. 4 and 30 m in the right lane at sampling No. 5, where the number of TMPs was 3–5. These numbers seem to be smaller than those from other studies (Mathissen et al. 2011; Sommer et al. 2018). Although it may be difficult to directly compare with such studies due to the difference in investigation techniques and the target size of the main focus (Foitzik et al. 2018), some possible factors are discussed in the following paragraph with a focus on road structure and driving behavior.

According to the Ministry of Land, Infrastructure, Transport and Tourism guideline (MLIT), it is required for the roadway to exhibit a gradient along the crossing direction. It was assumed that this gradient affects the distribution of TMPs, and TMPs in the right lane during rainfall were assumed to flow toward the left lane. The road structure suggests that when the weather is fine or rainfall is low, TMPs may remain in the right lane, or particles flowing from the right lane may not be able to flow toward the shoulder and may remain in the left lane. If the ease of flow toward the shoulder varies with the amount of traffic, this could explain why the number of TMPs smaller than 105 μm in the right lane was lower than that in the left lane.

The second factor, driving behavior, indicates that acceleration/deceleration increases within approximately 60–80 m based on the video survey, suggesting that TMPs are less likely to occur at 30 and 50 m because these distances are less affected by driving behavior.

Based on the above two factors, i.e., the road structure and driving behavior, it was assumed that the reason for the low number of TMPs smaller than 105 μm in the right lane is because no stopping behavior occurs at the 30 and 50 m points, and small TMPs tended to flow toward the left lane and shoulder due to the road structure.

In this study, road dust samples were collected on roads to better understand the actual conditions surrounding the occurrence of TMPs derived from tire dust. The findings are presented as follows.

• The number of TMPs tended to increase in proportion to the increase in traffic volume within the 10,000–30,000 vehicles/day range.

• The number of TMPs tended to increase immediately after rainfall and reached a plateau after approximately 4 days.

• Analysis of vehicle behavior suggested that the friction between the tires and road surface was significantly increased to over −10 m/s² of acceleration rate approximately 60 and 80 m from the stop line, which resulted in an increase in the number of TMPs of approximately 28% at approximately 70 m.

• The location where TMPs occur most frequently is influenced by the legal speed of the road.

This research was supported by the Environment Research and Technology Development Fund (JPMEERF20205R06) from the Environmental Restoration and Conservation Agency of Japan.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.