Tire wear mechanisms and how they relate to wear particle sizes – a brief review Open Access

Since the beginning of the 21st century, the pollution caused by microplastics (polymer particles between 1 µm and 5 mm) and its impact on the environment has become a major focus of scientific research (Browne et al., 2007). Among the various sources of microplastics, tire wear has emerged as one of the most significant contributors. Comprehensive studies conducted in Germany (Essel et al., 2015), Denmark (Lassen et al., 2015), and Norway (Sundt et al., 2014) have all identified dust from vehicle tires as the largest single source of microplastic emissions. In 2015 alone, approximately 6.4 million tonnes of synthetic rubber were produced globally for tire manufacturing (Boucher and Friot, 2017), and it is estimated that nearly one billion tires reach the end of their service life each year (Yadav and Tiwari, 2017). In the European Union, 3.2 million tonnes of used tires were generated in 2013 (CINARALP, 2015). Tire wear has been estimated to contribute approximately 5% to 10% of the total global microplastics entering the oceans (Kole et al., 2017), highlighting the urgent need for a better understanding and mitigation of microplastic pollution from this source. This is especially worrying as most particles end up in rivers, lakes or oceans, and have a profound negative impact on aquatic species (McCarty et al., 2023; Wagner et al., 2014; Chae and An, 2017).

Interest in tire wear as a source of particulate matter has intensified following the proposal of the EURO7 regulation (European Commission, 2024), which aims to establish limits on non-exhaust emissions (particles generated from vehicle tires and braking systems). The implementation of diesel particulate filters has already led to a reduction of over 99% in exhaust emissions of diesel vehicles within the PM₁₀ and PM_2.5 size fractions (Bergmann et al., 2009), resulting in minimal differences in exhaust emissions among electric, gasoline, and diesel-powered vehicles. Consequently, non-exhaust particles now account for more than 90% of total PM emissions, with approximately 10% attributed specifically to tire wear (Timmers and Achten, 2016).

This review aims to provide a comprehensive introduction to particulate matter, with a specific focus on particles generated through tire wear. We begin by discussing the fundamental characteristics of particles and proceed to examine the tribological mechanisms underlying tire wear. Experimental approaches and simulation methods commonly used to study these processes are also reviewed.

While existing literature on tire wear is extensive, most studies either focus on lab experiments to determine wear rates of rubber (Muhr and Roberts, 1992; Saibel and Tsai, 1969; Schallamach, 1954; Fukahori et al., 2020) or on collecting particles from roadways and analyzing their composition and size distribution (Zhang et al., 2023; Grigoratos and Martini, 2014). An additional challenge arises from the lack of standardization in this field. Tire abrasion can be examined from tribological, materials science or vehicle dynamics perspectives, each employing distinct analytical methods with different focal points. Even within a single discipline, variations in experimental design, parameter selection, and particle collection protocols hinder the comparability of results across studies. Therefore, a central objective of this work is to consolidate and critically assess the available data specifically related to theoretical and empirical predictions of tire wear particle sizes. Finally, we offer perspectives on emerging research directions and potential advancements in this field.

Tire wear particles form in the contact zone between the tire tread and the road surface. In general, tires consist of multiple materials that each perform a different function. The main materials constituting tires are listed, with sources, in Table 1.

From this variety of materials, natural rubber (NR) and synthetic rubber are prominent. The former gives the tire tread elasticity and flexibility and has good tensile strength, tear resistance and chemical resistance properties. The latter, e.g. styrene-butadiene rubber (SBR), improves hardness and has higher abrasion resistance than NR. Their combination forms a compound that provides the best properties of both materials.

Fillers are used as reinforcements to improve hardness, wear, and UV resistance, and softeners to improve the stickiness and wet grip performance. Steel and textile fibers increase stiffness, vulcanization agents accelerate the vulcanization process, and additives help to protect the tire against degradation by ozone, oxygen and heat (Piscitello et al., 2021).

In addition to their chemical composition, the size of tire particles plays a crucial role in their environmental and health impacts. Real-world data indicate that a significant amount of emitted particles fall within the size range of 2.5 µm to 10 µm (Lassen et al., 2015, p. 124). Particles in this range are linked to an increased risk of cardiovascular and cerebrovascular diseases, as they can trigger systemic inflammation, promote blood clotting, and even enter the bloodstream (Anderson et al., 2012). As a result, many field studies focus on measuring particles within this size range (Gustafsson et al., 2009; Aatmeeyata et al., 2009; Hussein et al., 2008; Kupiainen et al., 2005; Sjödin et al., 2010; Panko et al., 2009). However, some of these studies were not equipped to capture larger particles. Notably, other experiments using different methodologies have detected tire particles with size distributions centered around 20 µm to 80 µm (Kreider et al., 2010; Xu et al., 2025).

Tire wear is closely linked to the texture and quality of the road surface. Concrete and asphalt are the most commonly used pavement types. A typical concrete mixture consists of approximately 6% air, 11% Portland cement, 41% aggregate (crushed stone or gravel), 26% sand, and 16% water. In contrast, asphalt is composed of 95% aggregates bound together by bitumen (Ivel et al., 2020).

An exemplary particle size distribution is shown in Figure 1, where particles were collected on a plate, mounted on the frame of a car, immediately downstream of a tire (Dannis, 1974). The size and number of emitted particles were measured on a road with a concrete surface at velocities of 72 km/h and 105 km/h, and on an asphalt surface at 105 km/h. The results showed that driving on concrete produces more particles than on asphalt. The particle size appears to be log-normal distributed, where increasing velocity decreases the mean value, and the asphalt surface increases the standard deviation compared to the concrete pavement.

The impact of the road surface on tire wear was also investigated in (Allen et al., 2006), where tire wear was measured indirectly using chemical markers. The study was conducted in a tunnel that was initially paved with Portland Cement Concrete (PCC) and later resurfaced with Asphalt Rubber - Asphaltic Concrete Friction Course (AR-ACFC). Marker compounds present in tire tread material, exhaust gases, and aerosol samples were collected at the tunnel’s entrance and exit. Comparisons revealed that tire emission rates were 1.4–2 times higher on the PCC surface than on the AR-ACFC pavement.

The differences between porous (double-layered porous asphalt concrete, DLPAC) and dense (stone mastic asphalt, SMA) pavements were examined in (Lundberg et al., 2020). Porous pavement can reduce traffic noise and lower PM₁₀ emissions by trapping dust particles, however, it is less durable and therefore rarely used in Nordic countries, where studded tires are common. The study analyzed mineral dust loads (particles < 180 µm) and found that emissions peaked during winter months, decreasing by a factor of 4–5 in the summer for both pavement types. While the total particle emissions across the wheel tracks (left, right, and center) were similar for both DLPAC and SMA, the dense SMA pavement showed PM concentrations 8–10 times higher at the road edges. The particle size distribution is centered around 20 µm, with the DLPAC distribution slightly shifted toward larger sizes.

Gustafsson et al. (2009) focused on the impact of aggregate material and size in asphalt concrete, combined with different tire types (studded, winter and summer). The results showed that granite pavements produced approximately 70% more PM₁₀ than quartzite pavements of the same aggregate size. Between two quartzite pavements with different aggregate sizes, the one with a maximum size of 16 mm generated roughly three times more emissions than the one with a maximum size of 11 mm. In addition, studded tires produced about ten times more PM₁₀ emissions than standard winter tires, while summer tires contributed a negligible amount.

Another important factor in tire wear is vehicle weight, as tire abrasion is caused by friction between the tire and the road, which increases with the normal force exerted by the vehicle. Furthermore, resuspension of road particles is influenced by the size and aerodynamic properties of a vehicle, with larger vehicles generally generating more resuspension due to their greater mass (Timmers and Achten, 2016). Consequently, tire wear is generally assumed to be positively correlated with vehicle weight, a hypothesis that has been investigated and partially confirmed in several studies (Simons, 2016; Garg et al., 2000). This is especially relevant because the mass and front area of new car models have increased by 66% and 22% respectively, between 1980 and 2018 (Weiss et al., 2020). In addition, the weight of electric vehicles is on average 20% higher than their ICE counterparts (Mueller et al., 2024) and by 2050, 35% to 60% of all cars sold in the US could be electric (medium outlooks (Muratori et al., 2021).

While numerous factors influencing the size of tire debris particles have been discussed, the particle size is ultimately governed by the underlying wear mechanisms. In the existing literature, tire wear is typically categorized into three main mechanisms: cut and chip (CC) abrasion, fatigue or pattern wear, and chemical or smearing wear.

CC wear occurs when individual surface asperities or irregularities remove fragments of the tire tread, resulting in scratches aligned with the direction of sliding. Fatigue or pattern wear arises from repeated loading cycles on relatively smooth surfaces, leading to crack formation, culminating in wear patterns oriented perpendicular to the sliding direction. Chemical or smearing wear is driven by local heat buildup, shear stress, and environmental factors such as oxygen, ozone, and UV radiation, which chemically degrade the rubber compound.

Each mechanism is associated with a characteristic particle size range. Chemical wear can produce an oily residue (Gent and Pulford, 1983), while volatilization processes generate nanoparticles (Park et al., 2017). In contrast, CC wear typically results in larger particles, which may agglomerate with pavement debris to form conglomerates ranging from 220 µm to 1230 µm in size (Adachi and Tainosho, 2004). This finding was confirmed by an investigation in which particles were collected using an instrumented vehicle equipped with an on-board system capable of determining the mass-weighted size distribution (Truong et al., 2025). In that study, the authors attributed elongated and rough-textured particles to fatigue wear, with average particle sizes ranging from approximately 12 µm in suburban environments to more than 50 µm on rural roads. In addition, they identified smoother, rounder, and more irregular particles, which they suggested originated from local vaporization and subsequent condensation of tire compounds. These particles exhibited average sizes between 0.6 µm on highways and 2 µm on urban roads. In the following chapters, a detailed examination of each wear mechanism is provided.

If tires are subjected to harsh conditions (e.g., riding on a gravel road), they experience a wear phenomenon often termed CC wear (Stoček et al., 2021a). Cutting takes place when the tire hits a sharp object with enough force to penetrate the surface. Chipping then occurs because traction, breakage, or other forces cause the initiation of a tear in the rubber compound, usually at a 90° angle to the direction of the cut, removing a piece of tire thread (Beatty and Miksch, 1982).

Experimentally, CC can be investigated with an Instrumented Chip and Cut Analyzer (ICCA), where a rotating rubber sample is impacted by a stainless-steel tool with a specified frequency. After some revolutions, cracks with distinct angles form on the surface of the material, which grow in length and number, eventually intersecting, which leads to small chips of rubber being removed. The cracks in SBR appeared to grow deeper than in NR, creating larger wear particles. In contrast, the impact process progressed much more rapidly for the BR compound, leading to a higher abrasion rate and early material failure (Stoček et al., 2019).

The influence of the material on CC damage was further investigated with an ICCA using an impactor of 2.5 mm radius (Stoček et al., 2021b). At a lower normal force $≤$ 130 N, a 40 NR/60 SBR blend was more resistant, and at a higher force > 130 N, pure NR developed less damage. This is due to the higher intrinsic strength of SBR compared to NR, but at higher load, strain-induced crystallization occurs in NR, increasing its strength (Treloar, 1975).

For another material, hydrogenated nitrile rubber (HNBR), a linear relation was found between frictional work and the average particle diameter. This was tested by pressing a quartz sample (Mohs hardness 5.5–6.5) on a rotating rubber disk (modified Du Pont Abrader) (Thavamani and Bhowmick, 1992).

Numerical methods to describe abrasion at the asperity level are discussed in (Zhang et al., 2022), which gives some insight into the two classic theories of the Archard wear law and Rabinowicz criterion, the Finite Element Method (FEM) with realistic elastic-plastic material behavior, and molecular dynamics (MD) simulations.

In (Wu and Shi, 2013) FEM was used to confirm a previous theory (Suh, 1973) on how wear particles form in the contact of two metal asperities: First, the fracture initiates and propagates in the trailing portion of the contact area, followed by a second smaller fracture that emerges at the leading contact edge. Finally, the two fractures propagate and link below the contact area, forming a flake-like wear particle.

The contact of two asperities was also simulated using MD (Brink and Molinari, 2019). It was assumed that each asperity is hemispherical with diameter d and that when they touch, the formed junction interface is inclined by the angle $θ$ relative to the direction of movement. It was found that exists a critical $d*$ and $θ*$ for a certain strength of the adhesive interface that determine whether the said asperities slide past each other, plastically deform or adhere to each other and form a wear particle. The critical length scale $d*$ was also investigated in (Aghababaei et al., 2016), in which the authors found a function for $d*$ that accurately predicts if a debris particle is formed from the asperities or if they are plastically deformed and smoothed.

It should be noted, however, that all simulations were done with no specific material in mind (or with materials that behave like metals), and therefore, no effects unique to rubber were considered. Expanding and completing the simulation framework, including rubber-like materials, is an important research direction for the next decades.

In the end, the mechanism of wear is influenced by many factors, a fact nicely demonstrated by multiple experiments. When (Muhr and Roberts, 1992) abraded unfilled NR on silicon carbide paper in the presence of a lubricant, score lines in the direction of motion formed on the rubber surface, indicating CC wear. If no lubricant is applied, even though the CoF increased only slightly (1.34 vs 1.23), the score marks are replaced by a wear pattern. Similarly, an experiment on the aforementioned Du Pont abrader showed that NR and SBR formed a wear pattern, while in HNBR, fine scratch marks in the direction of abrasion are observed (again indicating CC wear). However, when the same experiment was repeated at 50°C, HNBR a wear pattern appeared as well (Thavamani and Bhowmick, 1993). The formation of a wear pattern perpendicular to the sliding direction indicates that a different wear mechanism is dominant: Fatigue.

It is well-known that rubber on a smooth surface does not slide continuously but in a stick-slip motion. This is caused by rubber molecules adhering to the surface, getting stretched along the direction of movement (stick phase), detaching and relaxing (slip phase), before they start to adhere to the surface again (Schallamach, 1963; Persson and Volokitin, 2006).

This was further explored by sliding a steel slider, of the razor blade type, over a block made from natural rubber, in reciprocal motion, while measuring the frictional force and the acceleration (Fukahori and Yamazaki, 1994b). Results show a stick-slip motion with a frequency $fss$ of 10 Hz to 20 Hz (decreasing with increasing normal force), where the frictional force increases during the stick phase and decreases in the slip phase in a periodic, sine-like manner. The acceleration is negligible in the stick phase, but shows violent micro-vibrations of a higher frequency (500 Hz–1000 Hz) during the slip phase. A further analysis revealed that this frequency of the micro-vibrations matches the intrinsic natural frequency of the rubber $f0$⁠, which was measured separately.

During micro-vibrations on the rubber surface, upward motion causes it to adhere to the slider. The resulting high relative sliding velocity generates frictional forces that induce a significant stress concentration at the edge of the small contact area, leading to the formation of the first micro-crack in the rubber. When the rubber surface detaches and subsequently readheres to the slider, a second micro-crack forms at a distance:

from the first, where v. is the sliding velocity and $f0$ is the natural frequency of the used rubber sample. This theoretical value was compared to the optically measured distance between the micro-cracks on the rubber sample. This relationship was confirmed for NR with multiple different velocities (Fukahori and Yamazaki, 1994b). The pattern spacing $d0$ then increases with the number of slidings, very similar to how small streams converge into a main river, until a critical value $d0=D0$ is reached, after which the pattern spacing stays constant and can be estimated similarly to equation (1):

using the stick-slip frequency $fss$ instead of the natural frequency. The formation of micro-cracks is sketched in Figure 2(a).

Even after a stable wear pattern has formed on the rubber surface (an exemplary wear pattern is shown in Figure 3), micro-cracks are continually formed on top of larger ridges, which creates wear particles in a size range from tens of micrometers or less up to hundreds of micrometers (Cadle and Wiliiams, 1978; Dannis, 1974).

The presented theory of pattern formation was also confirmed to apply to filled NR and SBR. Reinforcement by carbon black increases the stiffness of the material and therefore increases the initial and stick-slip frequency, which gives a smaller initial and final pattern spacing. Furthermore, the additional damping decreases the propagation speed of the surface waves, thus increasing the necessary number of cycles until a steady-state wear pattern is formed. Both effects lead to a smaller abrasive wear in more filled rubbers (Fukahori and Yamazaki, 1994a).

The underlying mechanism of fatigue wear is crack growth within the material. This is schematically shown in Figure 2(b). As the blade passes over the pattern, the tongue of rubber is pulled back and then released as the blade moves on. The stress produced by this process is assumed to cause crack growth (Southern and Thomas, 1978). Initially, the crack propagates at a low slope relative to the surface, until a critical length is reached, after which it turns upwards to form a new wear particle. When looking at fatigue crack growth from a fracture mechanics standpoint, it is therefore important to consider crack growth perpendicular to the surface. Experimental data suggest that, if the rate of growth is the same, then the whole time it takes to create the wear particle is more than twice the propagating time of the low slope (Uchiyama and Ishino, 1992; Southern and Thomas, 1978).

Notably, stick-slip motion is not a unique phenomenon to rubber friction but can occur in any sliding contact if the difference between the static and dynamic coefficients of friction is large enough. If the material is considered as a damped harmonic oscillator, the transition from a stick to a slip phase acts as a displacement. Therefore, subsequent vibration in the (damped) natural frequency is very much expected (Bowden and Tabor, 2001; Gao Chao et al., 1994). This could also explain the influence of the normal force on the stick-slip frequency, because the load directly influences the friction coefficient of rubber (Fortunato et al., 2017).

The linear rate of abrasive wear $D˙$ can be estimated from the crack growth rate $dc/dn$ if the mean strain $ϵ*$ in the stress field, produced by frictional sliding, is accurately evaluated, see equation (3) (Fukahori and Yamazaki, 1995).

Once the wear pattern stabilizes in size, and assuming perfect stick during the stick phase, the mean strain amplitude can be expressed as follows:

where $Lst$ and $Lsl$ are the distances the slider moves in the stick and slip phase, respectively, $μ$ is the friction coefficient, P is the normal load, E is the Young’s modulus of the rubber, and S is the deformed cross-sectional area, approximated by $S=hd$⁠, with h being the width of the rubber specimen and d the indentation depth of the slider. The unique combination of properties of rubber materials, high friction, low modulus, and high deformability, is the reason why periodic surface patterns form on rubber but rarely on other materials. In reality, however, perfect slick does not take place even in the stick phase (also micro-vibrations are still generated), and the slider moves a certain distance $Δr$ in the stick phase. The formula for the mean strain can be adapted to accommodate this:

The actual and ideal mean strain values calculated match perfectly with the authors’ experiments (Fukahori and Yamazaki, 1995). The results also show that the rate of crack growth is not improved by the use of carbon black as a filler agent, at low strain rates, it is even faster than in unfilled rubbers. However, as the carbon black content increases, E increases, $μ$ increases slightly, and S decreases, which according to equation (4) drastically decreases the strain $ϵ*$ and therefore the wear rate per equation (3).

Earlier experiments on pattern wear of rubber with sandpaper have already shown some proportional relations [equation (6)] between the pattern spacing $D0$⁠, the normal load per unit area L, the curvature of the abrasive particle r, the diameter of the abrasive particle d, and the Young’s modulus E (Schallamach, 1954):

If the abrasive can be approximated to a set of close-packed hemispheres (as a simplified model of an asphalt surface), the following relation can be made:

This linear increase in additional wear due to pattern formation with pattern spacing was also observed in other experimental data (Schallamach, 1968a, 1968b). Other results (Southern and Thomas, 1978) found that pattern spacing is directly proportional to the frictional force (per unit width).

The equations presented in this chapter should be applied with caution, as rubber wear is a highly complex phenomenon influenced by numerous parameters, as previously discussed at the end of Chapter II. For instance, different rubbers can exhibit varying wear pattern spacings even when subjected to similar modulus and load conditions (Southern and Thomas, 1978). There is evidence suggesting that the proximity to the glass transition temperature (⁠$Tg$⁠) plays a significant role. In the case of NBR (⁠$Tg$ = −30°C), the wear pattern is not pronounced at 20°C, but becomes more distinct as the temperature increases. Conversely, rubbers such as NR and SBR, with lower glass transition temperatures (⁠$Tg$ = −55°C and −70°C, respectively), exhibit a much coarser pattern at 20°C, which tends to become finer with increasing temperature.

The coefficient of friction and the presence of lubrication also significantly affect abrasion behavior, as described by Eq.4. In one study (Setiyana et al., 2021), tests conducted under identical parameters, but one dry and one with a thin layer of low-viscosity oil, showed no meaningful difference in pattern spacing, consistent with the observation that the measured friction coefficients were nearly identical. However, other results (Muhr and Roberts, 1992) demonstrated that the wear pattern can completely disappear when a lubricant is applied, due to the reduced frictional forces. Additional experiments (Muhr et al., 1987) indicated that friction force may not be the dominant factor, as the wear rate was found to decrease by several orders of magnitude, which is far more than the reduction in the friction coefficient alone would suggest. Furthermore, evidence of strongly nonlinear behavior has been reported (Evstratov, 1967), with the wear rate dropping sharply when the friction coefficient falls below a critical threshold.

The formation of a wear pattern is prevented if the direction of the abrader is reversed after every run. Wear under these conditions is called intrinsic abrasion, and it is closely connected to the tensile strength of the rubber material. The abrasion rate and the size of abraded particles, created under these circumstances, are significantly lower (Schallamach, 1954; Schallamach, 1958; V H Nguyen et al., 2017).

Directional effects were investigated using the Grosch wheel experiment, in which a small rubber wheel is pressed against a rotating abrasive disk (V H Nguyen et al., 2017). By adjusting the wheel’s orientation, different slip angles relative to the rolling direction were imposed. The slip angle was periodically varied under three conditions: (a) +45° to −45°, (b) +45° to +60°, and (c) held at constant 0°. The results revealed that the wear rate was the highest for the largest slip angle variation, following the trend: (a) > (b) > (c). Remarkably, the accumulated material damage exhibited the opposite trend, with the 0° slip condition, which corresponds to pattern wear, causing the greatest damage. Over time, however, differences in both wear rate and damage diminished, converging toward similar steady-state values across all slip angle conditions.

When NR is abraded against a smooth surface, such as finely ground glass, the temperature rise due to frictional heating, $ΔT$⁠, has been shown to scale with the sliding velocity v, normal force F, and a material-dependent exponent $α$ (Schallamach, 1957):

This frictional heating softens the rubber and increases the real contact area, generally leading to a decrease in the CoF with increasing normal load (Schallamach, 1955; Fortunato et al., 2017). The CoF also exhibits a nonmonotonic dependence on sliding velocity, showing a Gaussian-like peak centered around v ≈ 1 mm/s, spanning several orders of magnitude. This behavior is observed on both smooth and rough (e.g., asphalt) surfaces. Temperature rise under frictional loading has also been studied using a thermocouple sliding against a rotating rubber disk (Schallamach, 1958). Results show that temperature increases more rapidly at lower normal loads, attributed to reduced heat conduction due to less flattening of surface asperities. Consequently, softer materials tend to exhibit lower surface temperatures during friction, despite lower thermal conductivity.

It may be surprising that fatigue, a phenomenon associated with blunt tracks, can be investigated by scraping with a razor blade. However, the sharpness of the blade has no influence on the wear pattern formation, as long as it is sufficiently sharp so that only one element of the wear pattern is deformed at a time (Southern and Thomas, 1978). Furthermore, lubricants drastically decrease the rate of blade wear, as explained earlier, even if the mechanism is not fully understood. These two findings indicate that, especially for a dry contact, frictional stress, rather than cutting, is primarily responsible for the wear (Muhr and Roberts, 1992). Also, while many experiments about pattern wear are carried out on a blade abrader, wear patterns are observed in real tire-road contact (Schallamach, 1958; Grosch and Schallamach, 1961).

If the normal force, and consequently the frictional energy, are sufficiently low, a sticky layer can develop on the rubber surface instead of a typical fatigue wear pattern. This phenomenon is known as chemical wear (M Huang et al., 2018).

Under relatively mild wear conditions, certain rubber materials (CB-filled compounds of SBR, NR, and EPR) exhibit the formation of an oily, sticky and degraded surface layer, a phenomenon known as smearing wear (Wu et al., 2023). This wear debris is characterized by a higher concentration of volatile components [11.12% vs 8.13% in NR, and 10.23% vs 8.75% in SBR filled with 50 phr CB (parts per hundred rubber)] as well as elevated oxygen content (41.7% vs 39% in NR and 8.5% vs 2.8% in SBR with CB). Although $Tg$ remains similar, the molecular weight is about an order of magnitude lower than that of the bulk material, accounting for its sticky, liquid-like morphology.

A significant de-crosslinking effect has also been observed (Wu et al., 2023), with reductions of approximately 75% in NR and 55% in CB-filled SBR. This suggests thermo-mechanical degradation of the three-dimensional rubber network. Evidence of devulcanization is further supported by solubility tests in toluene: Freshly generated smearing wear debris is completely soluble, while samples exposed to air for over 24 h become insoluble. A likely mechanism involves the formation of free radicals under loading conditions – arising from the scission of polymer backbones or crosslinks, followed by radical scavenging facilitated by filler particles and antioxidants.

This behavior has been corroborated by previous studies (M Huang et al., 2018), which additionally report that aged smearing wear debris can be redissolved following brief sonication. This suggests the formation of bound rubber, where the polymer interacts with filler particles through relatively weak physical or chemical associations.

This microstructural phase separation within a polymer is illustrated in a series of SEM images in Figure 4 (Koliolios et al., 2024). The leftmost image provides an overview of the fractured wear surface, revealing a flake-like texture indicative of layered material failure. A zoomed-in section in the middle panel offers a closer view of the microstructure, highlighting heterogeneous regions within the material. Further magnification on the right distinguishes two distinct phases: The upper inset shows a dark, textured region where carbon black particles are densely clustered and closely interact with the polymer matrix, resulting in a more rigid, constrained phase. In contrast, the lower inset reveals smoother, layered structures representing polymer areas with little to no filler interaction. This clear phase separation supports the conclusion that smear wear layers exhibit weaker mechanical properties due to the loss of uniform filler dispersion and the emergence of mechanically distinct polymer domains.

The formation mechanism of smearing wear was long not fully understood (Gent and Pulford, 1983), however, oxygen appears to play a significant role in the process. This was confirmed by conducting an abrasion experiment with the same material under the same conditions, only once in an oxidative environment, and once in an oxygen-free atmosphere (Nakano et al., 2021). In the absence of oxygen, the rubber produced a powdery substance, while in the presence of oxygen, smearing wear occurred.

This oxygen-dependent behavior was also observed in a similar experiment (Gent and Pulford, 1979), and the abrasion rate increased eightfold under oxidative conditions. Comparable wear rates were recorded in inert environments such as nitrogen, argon, or vacuum, indicating that the absence of reactive species prevents smearing. Interestingly, when a free-radical-trapping compound, such as thiophenol, was introduced under an inert atmosphere, the formation of oily smearing wear was again observed. These findings suggest that the critical factor in smearing wear formation is not oxygen per se, but the presence of a radical-trapping agent.

The stability of polymeric radicals is closely correlated to the rate of metal wear of the knife abrader. A highly reactive radical is thought to primarily react internally and thus cause little wear to the scraper, whereas more stable polymer radicals attack the metal more. For example, radicals formed by NR and SBR are particularly stable in nitrogen, but readily react with oxygen in air to produce smearing wear. EPR, on the other hand, undergoes degradation to a liquid-like state by wear in air and in nitrogen, because in either case, the free radicals formed by molecular rupture undergo hydrogen-abstraction and disproportionation reactions, causing further decomposition of the polymer (Baldwin and Strate, 1972; Gent and Pulford, 1979). BR is highly reactive and therefore prefers to react internally with other polymer molecules, which can even increase crosslinking within the polymeric matrix, creating only dry debris. Finally, butyl rubber [Poly(isobutylene-co-isoprene)] is the only tested material that experiences roughly three times more wear in air than in nitrogen. This is explained by the relatively stable macro-radicals formed by the polymer itself, but the even more stable peroxy radicals which are formed with oxygen (Gent and Pulford, 1979).

Another way to influence whether and to what extent smearing wear occurs is by feeding dust into the contact zone (Schallamach, 1968a, 1968b). This was tested on an Akron Abrasion Tester, where a smaller rubber sample wheel rubbed against an abrasive wheel. The dust was either a powder mixture of silicon carbide and Fuller’s earth or magnesium oxide. The results are shown in Table 2. It was demonstrated that for SBR when using magnesia, the wear under an air atmosphere is 4 to 5 times higher than under a nitrogen atmosphere. In contrast, when the powder mixture is used, the wear is slightly higher under a nitrogen atmosphere. The given reasons are that magnesium oxide has shown lubrication properties (Grosch K A, 1963) and is an effective adsorber of smearing wear. Fuller’s Earth, on the other hand, is a poor adsorber and has a particle size some 70 times larger than magnesia, which makes it a better abrasive. However, the performance of magnesia as a solid lubricant has only been evaluated in air (Scharf and Prasad, 2013), and its lubricating properties may not be transferable to other environments. Nevertheless, the presence of magnesia as a lubricating agent could account for the observed reduction in abrasion rates. An open question remains: If magnesia inhibits the friction-reducing effect of smearing, why is the abrasion rate in air still higher than in nitrogen? One possible explanation is that, in air, the smeared wear material is continuously formed but immediately re-adsorbed, preventing the development of a stable, protective tribolayer.

It is now evident that oxygen plays a crucial role in the formation of smearing wear. Consequently, the presence of antioxidants (AOs) in the rubber compound may also significantly influence this process. Indeed, as shown in Table 2, natural rubber compounds without AOs exhibit abrasion rates approximately 50% higher in air compared to those containing AOs, a trend that is consistently observed across all experimental conditions.

A different study using a blade abrader in air (Pulford, 1983) investigated three CB-reinforced NR compounds, two of which contained different AOs. Under low frictional force, the NR compound without AO exhibited approximately three times higher wear compared to those with AOs. As the applied force increased, the nature of the wear debris changed: at a certain threshold, it transitioned to a powdery form, suggesting the onset of a different, more abrasive wear mechanism.

Another noteworthy observation from Table 2 is that, for SBR, the influence of the surrounding atmosphere on abrasion behavior appears to be less pronounced when tested against a rough surface (grinding wheel), which is also supported by other findings (Uchiyama, 1986).

Measuring the impact of ozone was done by comparing the abrasion of EPR, which is quite resistant to ozone, to NR with an AO. No noticeable difference in abrasion can be seen in Table 2. Also, fragmentary experiments were carried out in air that passed through a saturated potassium iodine solution to break down any ozone, and again, no change in the abrasion rates was measured. This could indicate that oxygen, rather than ozone, is the driving factor in the formation of smearing wear.

The effect of humidity on the abrasion rates was also investigated (Schallamach, 1968a, 1968b). In general, wear increased more for NR with no AO than for NR with AO or SBR, and even decreased for polybutadiene rubber (BR). The measured difference increased with rising humidity. Part of the effect could also be due to the trivial fact that more dust sticks to the sample in a humid atmosphere, but this would not explain the result of the BR compound.

The effect of debris accumulation in the wear zone was investigated by (M Huang et al., 2018) using a tribometer. Their results showed that longer wear cycles led to greater accumulation of debris on the wear track, which in turn corresponded to a reduction in the overall wear rate. The study concluded that smearing wear generally exerts a protective effect against further abrasive wear by forming a layer that partially shields the underlying rubber surface.

Chemical degradation of tire tread can also lead to the formation of nanoparticles through volatilization processes. In a study by (Park et al., 2017), tire samples were heated in a reaction chamber, with a controlled airflow directing the emitted particles to a measurement system. A significant increase in particle generation was observed once the tread temperature exceeded 160°C, with particle concentration rising exponentially up to 350°C. The resulting particle size distribution was unimodal and followed a log-normal profile, centered between 60 nm to 100 nm. While higher temperatures increased the number of particles without altering the size distribution, slower cooling rates produced larger particles. Unlike the elongated, sausage-like morphology commonly observed in road wear particles(Adachi and Tainosho, 2004; Kreider et al., 2010), the particles formed in this study were nearly spherical and composed primarily of carbon, oxygen, silicon, and sulfur – elements derived from the rubber compound. These findings support a formation mechanism involving the volatilization of organic compounds in the tire tread due to frictional heating, followed by gas-phase condensation into nanoparticles.

This work provides an overview of the three primary tire wear mechanisms, outlining the conditions under which they occur and their relationship to the formation of wear particles, see Table 3.

Abrasion on a rough surface and under high normal pressure results in a high tearing energy. Under these conditions, cut-and-chip wear occurs, where particles are ripped from the bulk rubber material. It was found that, at least for an HNBR disk abraded by quartz, the particle size is proportional to the tearing energy.

On smoother surfaces, where tearing energy is lower, fatigue wear becomes the dominant mechanism. The stick-slip motion of rubber against the surface creates a stress field that promotes crack propagation, forming characteristic wear patterns. It is assumed that the size of the resulting particles is related to the spacing of these patterns, which is influenced by factors such as applied load, coefficient of friction, Young’s modulus, the counterbody surface, temperature, and possibly other yet unidentified parameters.

At low normal loads, tearing energy is minimal and chemical wear dominates. In this regime, a sticky, oily film forms on the rubber surface, composed of oxidized and de-crosslinked polymer chains. This layer can act as a lubricant, generally reducing wear rates. However, at temperatures above 160°C, nanoparticles may form through volatilization processes.

It is worth noting that much of the foundational research into tire wear was conducted over 30–70 years ago. These studies often used experimental setups that are now considered outdated. Moreover, different experimental methods, parameters, and protocols make direct comparison of results difficult, which consequently can lead to results that seem inconclusive or even contradictory.

With advances in materials science, imaging techniques, and computational modeling, there is a significant opportunity to revisit tire wear phenomena using modern methods. These developments open the way for future studies that combine tribological testing with particle characterization, thereby enabling a deeper understanding of tire wear processes and the underlying mechanisms. Emerging research efforts already illustrate this potential, for example: standardizing tire wear particle emission studies (Mennekes and Nowack, 2022); using finite element simulation software, with data from sensors embedded in tires, to feed a neural network-based tire wear algorithm (Li et al., 2021); focusing on environmental damage in other domains such as soil (Ding et al., 2023); or evaluating the toxicity of certain tire compounds, e.g. the antioxidant 6PPD (Chen et al., 2023).

A deeper understanding of the interplay between tire wear mechanisms will not only enhance tire durability but also inform strategies for reducing the environmental impact of tire-derived particles.