Quantitative study of triboemission kinetics from polymer fiber-reinforced mortar paving blocks: Unravelling the dynamics of nanoparticle aerosol release

Quantitative study of triboemission kinetics from polymer fiber-reinforced mortar paving blocks: Unravelling the dynamics of nanoparticle aerosol release F. Husanu a , ´ A. Alonso b , V. Calder´ on b , M. Castellote a , R. Nevshupa a,* a Eduardo Torroja Institute of Construction Sciences, Spanish National Research Council (IETCC-CSIC), C/Serrano Galvache 4, Madrid 28033, Spain b Department of Architectural Constructions and I.C.T., Higher Polytechnic School, Burgos University, C/Villadiego, s/n, 09001 Burgos, Spain ARTICLE INFO Keywords: Mortar paving blocks Waste additives Recycled polyurethane fibers Nanoparticle aerosols ABSTRACT Triboemission of nanoparticle aerosols from construction materials is a growing concern due to its potential impact on air quality and human health. In this study, we investigated the effect of aggregation of polyurethane fibers (PUFs) proceeding from waste on the kinetics of triboemission in cement mortars. A quantitative methodology was employed to assess the deposition rate, particle size distribution, and emissivity for the aerosols within the particle aerodynamic diameter range of 10–400 nm. The triboemission properties were correlated with the pore structure, morphology and tribochemical transformations of the particles and worn surfaces. Our results highlight the intricate influence of PUF aggregation on the kinetics of triboemission in cement mortars through both direct and indirect mechanisms and provide valuable insights into the mechanisms governing triboemission in construction materials. 1. Introduction The evolution of internal combustion engines and the growing adoption of electric vehicles in urban settings have significantly changed the composition of emitted pollutants. This shift has resulted in a notable rise in the prominence of fine and ultrafine particulate matter (PM) [1–6]. Recent studies have highlighted the substantial contribution of particulate matter pollution to worldwide attributable deaths, ranking it as the third biggest killer, accounting for 11 % of such deaths [7]. Martins et al. [8] highlighted that decreasing exposure to fine and ultrafine particles is a complex task that necessitates a thorough understanding of the particles’ origin. Among the sources of non-exhaust emissions, the release of PM into the atmosphere from the wearing and weathering of tires, brakes, transport infrastructure, and pavement materials has emerged as a significant contributor. Moreover, in the context of the circular economy, the need for effective control of PM emissions has become even more crucial as the incorporation of recycled materials becomes increasingly prevalent. The circular economy framework is built upon three pillars: waste and pollution reduction, product and material reuse and recycling, and ecological restoration. Cementitious products are widely used for their ability to incorporate and reuse by-products or wastes from diverse industries [9–13]. Recent studies highlighted that polyurethane, fibreglass, polyester, epoxy or furan fibers added to mortar can improve its functional, mechanical and bonding properties [14–17]. Several studies explored the addition of waste polymeric materials such as tire rubber particles, as partial substitutes for fine aggregates in both normal-strength and high-strength cement concrete [18,19]. These studies have revealed that the inclusion of polymer aggregates can have contrasting effects. For example, crumb rubber aggregates have been found to improve concrete’s resistance to abrasion but decrease its compressive strength, whereas the addition of polypropylene fibers enhances both the tensile and compressive strengths of concrete. According to Corinaldesi et al. [20] the inclusion of 10 % polyurethane, styrene-butadiene rubber waste or scrap derived from discarded rubbershoe outsoles led to a notable decrease in the compressive and flexural strength of the mortars due to increased porosity. The modification of structure and mechancial properties of mortars by the additives can influence the aerosol generation during machining or fragmentation of construction materials and lead to a significant increase in air pollution [21], which can overshadow the environmental benefits of recycling construction materials. Numerous studies have been conducted to investigate the size distribution, morphology, and chemical composition of nanoparticle (NP) aerosols emitted from * Corresponding author. E-mail address: [email protected] (R. Nevshupa). Contents lists available at ScienceDirect Cement and Concrete Research journal homepage: www.elsevier.com/locate/cemconres https://doi.org/10.1016/j.cemconres.2024.107650 Received 13 July 2023; Received in revised form 16 July 2024; Accepted 21 August 2024 Cement and Concrete Research 185 (2024) 107650 Available online 27 August 2024 0008-8846/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). construction materials during sawing, sanding, drilling, cutting [22–24], and regular wear [25–29]. These studies have primarily focused on analyzing the concentration time series of emitted particles, which are highly influenced by the specific configuration of the experimental setup and test conditions. However, limited research [30] has been dedicated to quantifying the absolute kinetic parameters associated with PM emissions in abrasion, wearing or fragmentation of construction materials. Heichelheim et al. [31] studied aerosol generation in fragmentation of concrete with various microand nanoadditives and found that the relative size and type of inclusion significantly affects the size of particles produced when fragmented. The inclusion effect on particle size is different for the macro-scale fragments and the nano-particles aerosols. For example, slag concrete acted more uniformly when compared with fly ash concrete, producing a greater proportion of large fragments. Steel fiber-reinforced concrete produced smaller particles than the other mixes studied in [31]. The influence of the additives incorporation on aerosol generation can occur through two primary mechanisms. Firstly, additives can be directly released into the aerosol through mechanisms such as detachment from the abraded material, fragmentation or tribochemical decomposition [27,32]. Secondly, additives can indirectly influence aerosol generation by modifying the intrinsic properties of the bulk material, such as its strength, plasticity, or porosity [22,33,34]. However, the relationship between these factors and aerosol emission is not always straightforward. Studies have shown that increased porosity can lead to either a decrease in aerosol emission rate, as observed during wood abrasion [35], or an increase, as observed in the abrasion of polymers, minerals, and composites [33,34]. Given the significant influence of mechanical properties and porosity on triboemission behavior, it is necessary to investigate the complex physical and chemical interactions of the recycled polyurethane fibers (PUFs) with the mortar and the effect of their addition on the emission of fine and ultrafine aerosols in abrasion. This research is expected to provide the groundwork for the establishment of a standardized methodology to evaluate the aerosol emissivity of solid materials, with a specific focus on cementitious construction materials. By addressing this crucial aspect, our study aims to contribute to the development of consistent and reliable assessment methods in the field, thereby enhancing our understanding of aerosol emissions associated with these materials. 2. Materials and methods 2.1. Materials Eco-friendly paving blocks were developed using recovered waste materials. These materials consisted mainly of recycled polyurethane fibers (PUF) obtained from shredded car interior linings and ceiling materials. The PUF also contained a smaller amount of polyester and glass fibers. The mortars consisted of CEM I 52.5 R Portland cement, washed natural 0/4 mm sand, and superplasticizer additives. The cement-towater ratio was consistent for all mortar compositions, while a portion of the sand was replaced with an equivalent volume of shredded PUF. Three different mortar compositions, labelled as A, B, and C, were prepared, each with varying concentrations of PUF. The PUF substitution levels were set at 20 %, 40 %, and 60 % of the sand volume, respectively. A benchmark mortar without any added PUF, referred to as Ref., was also produced. All samples were cured for 28 days. Further details regarding the specific fabrication process and characterization of the samples can be found in a separate publication [36]. 2.2. Triboemission measurements The experiments utilized a custom-designed pin-on-disk tribometer equipped with an enclosed aerosol-tight chamber (Fig. 1). This chamber has a volume of 14.57 l, uncertainty due to a complex geometry of the tribometer u (V) =0.35 l (Fig. 1). The tribometer consisted of a turntable connected to a motor shaft, a sample holder installed on the turntable, a spherical alumina abrader 20 mm in diameter, and a system for applying a constant load of 1.3 kg to the abrader. Sliding velocity was constant at 1 m/s, while the rotating speed was set at either 478 or 637 rpm depending on the diameter of the wear track, which was 40 mm or 30 mm, correspondingly. The size-fractionated aerosol concentrations were measured in the range 10–400 nm of electrical mobility particle diameters using a Scanning Mobility Particle Sizer (SMPS) connected to the aerosol-tight chamber using a stainless steel pipe. The airflow rate, Q 0 , at the inlet of the SMPS was set to 0.750 l/min. The tribometer was installed in a laminar flow extractor hood (HEPA EU14) to avoid the influence of ambient aerosols on the measurements. The concentration of ambient airborne particles, C 0,i , inside the environmental chamber was maintained below 30 #/cm −3 in any size range. For microscopy analysis the aerosol particles were sampled during additional test runs using the method suggested by R’Milli [37]. An original method based on the analysis of concentration time series [30] was used to quantify the triboemission rate of aerosol particles. This method is a further development of Koivisto’s [38] mass-balance method and its adaptation to the mechanical generation of aerosols. The deposition velocity was assessed from the analysis of the exponential concentration decays which occurred after the end of abrasion. The following solution was employed to determine the instant triboemission rate of aerosol particles within the i-th size range [30,38]: Kte,i(t) = V(dCi dt +Ci(t) τ 2,i)−Q0C0,i(1) where τ 2,I is the time constant of the exponential decay of aerosol concentration (i-th size range), C i (t) is the instant aerosol concentration. The total yield of the emitted aerosol nanoparticles during abrasion, Y te,i , was determined by integration of (1) over the time interval between the abrasion beginning (t =t 0 ) and the end of transient concentration decay, t b . Since Ci(t) |t0<t<tb≫C0,i&Ci(t0) = Ci(tb), the total yield can be found from the simplified expression: Yte,i=V τ 2,i∫tb t0 Ci(t)dt =∫tb t0 Kte,i(t)dt (2) The deposition rates, A d,i , were determined from the time constants of the exponential concentration decays: Ad,i=V τ 0− τ 2,i τ 0 τ 2,i(3) where τ 0 is the time constant of gas exchange between the aerosol Fig. 1. Schematic representation of the experimental setup. Adopted from [30] under Creative Common CC BY license. F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 2 chamber and the environment and τ 2,i is the empirical time constant of the concentration decay. The τ 0 was experimentally determined using the method of labelled gas, which yielded the τ 0 =1050 s (u( τ 0) =3.18 s) (see Supplementary Materials for more details). 2.3. Surface and material characterization The volume of the worn material was determined from the 3D measurements of the wear tracks obtained using confocal white-light microscopy (see Fig. S1 in Supplementary Materials). At least 15 measurements of cross-section area along the wear track were averaged for each track and then multiplied by the track length. The skeletal density of the mortars was evaluated using He pycnometry, while the apparent density was determined as the ratio of the mass of parallelepiped-shaped mortar samples to their macroscopic volume. The assessment of pore structure involved the utilization of dynamic water vapor sorption (DVS) methodology. Prior to the experiment, calibration of the DVS system was performed following the standard procedure ASTM E2551, using microcrystalline cellulose particles with sizes ranging between 20 μ m and 300 μ m. The water adsorption measurements were conducted at a temperature of 25 ◦C. The relative humidity (RH) was incrementally increased from 3 % to 90 % in 10 steps, followed by a gradual decrease back to 3 % in 9 steps. Each step allowed for a stabilization period of three hours to attain mass equilibrium. Additionally, an N 2 adsorption test was conducted at the boiling point of liquid nitrogen, following the standard procedure ASTM D4222–20. The obtained results were analyzed using the BrunauerEmmet-Teller (BET) model of multilayer adsorption to derive relevant pore structure parameters. The chemical characteristics of the worn surfaces were analyzed using Raman spectrometry using a laser with a wavelength of 780 nm and compared with the pristine (unworn) surfaces of the same samples. The contact angle between water and mortar surfaces was determined utilizing the sessile drop method. Drops of deionized water, ranging in volume from 1 to 4 μ l, were employed for the measurements. 3. Results and discussion 3.1. Triboemission of aerosols Fig. 2 presents a typical concentration time series for sample A, focusing on the particle size bin centered at 86.6 nm. This serves as an example of the experimental data obtained during the abrasion tests. As the abrasion progresses, the aerosol concentration initially rises. The peak concentration is reached either in the middle or towards the end of abrasion, depending on the specific mortar type. After the abrasion process stops, the concentration follows an exponential decay, gradually decreasing back to the background level. Fig. 3a displays the time constant values of this exponential decay, plotted against both particle size for various mortar types. Fig. 3b shows the corresponding deposition rate constant calculated using (3). A consistent trend is observed across most particle sizes, with the exception of the 20–60 nm range. For sizes outside this range, the time constant generally increases with particle size. This correlates with the decreasing trend of the deposition rate constant (A d ) in Fig. 3b. This can be explained by the diminishing influence of Brownian diffusion on larger particles, which becomes the dominant deposition mechanism in laminar airflow conditions [39]. However, within the 20–60 nm range, an interesting deviation from the trend is observed depending on the mortar type. Mortar A exhibits a similar pattern to the reference mortars, while mortar B shows a flatter trend, and mortar D displays an inverted U-shaped curve. This suggests that the presence of recycled Fig. 2. Concentration time series for sample A during the triboemission experiment. The plotted signal corresponds to the bin representing a mean particle size of 86.6 nm. The red line represents the exponential fit of the experimental data. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 3. a) The time constant values of the exponential decay as a function of both particle size and mortar type; b) the deposition rate constant as a function of both particle size and mortar type. F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 3 polyurethane fibers (PUFs) in the mortar significantly influences the deposition behavior of particles within this specific size range. To gain a deeper understanding of how particle size affects emission, we analyzed the instantaneous emission rates, which represent the rate at which particles are released during abrasion. Fig. 4a, d, and g show the emission rates as a function of nanoparticle diameter for the reference sample, samples B, and C, respectively. The emission behavior of sample A was very similar to sample B and is therefore omitted for clarity. In all cases, emission began as soon as abrasion started, with some minor fluctuations, and stopped abruptly when abrasion ended. Interestingly, mortars A, B, and the reference sample exhibited peak emission rates at the beginning of the test. However, mortar C displayed a contrasting pattern, with the emission rate gradually increasing throughout the abrasion process. This variation in emission trends observed across different studies highlights the lack of a universally accepted model to explain them. For instance, Bressot et al. [40] reported a decreasing linear trend in nanoparticle emission during the abrasion of masonry bricks, while Shandilya et al. [32] found a constant emission rate. In our study, we believe the observed differences in emission trends can be attributed to the influence of PUF aggregates on the mechanical properties and wear resistance of the mortars (further discussed in Section 3.4). We further analyzed the size distribution of aerosol particles at both the peak and steady emission stages (denoted as ν 1 and ν 2) using histograms presented in Fig. 4c, f, and i. No significant differences in size distribution were observed between samples A and B. However, for mortar C, a slight increase in the fraction of ultrafine particles with a mean diameter of 27.4 nm was identified during the peak emission phase. The reference mortar displayed a more interesting trend. The peak emission intensity shifted in size: during the initial phase, it was centered around 154 nm, while during the steady phase, it moved to around 116 nm. This shift could indicate changes in material properties or emission mechanisms during abrasion. Despite these variations, the average normalized particle size distributions (refer to Fig. 4j) were similar for all fiber-loaded and reference samples. These distributions were well-represented by a bimodal lognormal function, with the dominant mode centered around 154 nm. These findings are consistent with previous research [21,28,41–43]. The total amount of particles (within the 10–400 nm size range) emitted during abrasion was calculated. This value, called the cumulative emission, Y n,Σ , was obtained by integrating the emission rate over time (using Eq. 2) and summing the values for all particle sizes. The results are displayed in Fig. 5a with circles. The corresponding volume of material worn away during abrasion, V w , is represented by the triangles. Fig. 5a reveals an interesting trend. The cumulative emission of particles exhibits an asymmetric U-shaped curve in response to the concentration of PUFs in the mortar. In contrast, the worn volume shows a more complex relationship with PUF concentration. While sample A exhibits a significant increase in worn volume, samples B and C (with higher PUF content) show a noticeable reduction (33 % to 55 %) compared to the reference sample. This decrease in worn volume for samples with higher PUF content aligns with previous research [18], which demonstrated that incorporating waste tire rubber particles into concrete improves its abrasion resistance. The observed non-linear behavior in our study can likely be attributed to two opposing effects of PUFs on the cementitious matrix. On one hand, the PUFs may act as a form of reinforcement, strengthening the material. On the other hand, they might also increase the porosity of the mortar. These competing effects will be discussed in more detail in Section 3.4. Considering that wear is the predominant emission mechanism, this study introduces a new parameter - the normalized or specific yield of emitted particles per unit of worn material - which can be calculated as the ratio: Fig. 4. Time series of emission rates as a function of particle size for reference sample (a), samples B (d), and C (g). Average time series for the selected range of particle diameters denoted h 1 for reference sample (b), sample B (e), and sample C (h). The histograms of size particle distributions during intensive and steady emission for reference sample (c), sample B (f), and sample C (i). The distribution function of particle size for all samples averaged over the entire test duration (j). F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 4 NV=Yte,Σ/Vw(4) The normalized yield reflects a material’s ability to emit aerosol particles, or its “emissivity,” during abrasion under specific conditions. The N V exhibits a V-shaped trend with a minimum point observed around a PUF concentration of 20 %. Interestingly, sample B displays a similar emissivity to the reference sample, despite exhibiting significantly higher wear resistance. The factors contributing to this behavior will be further explored and discussed in Section 3.4. It’s important to acknowledge that normalized emission is just one parameter derived from triboemission research. Several other parameters can be introduced depending on the research goals. For instance, in the study of aerosol emission from pavement the specific yield of emitted particles can be related to factors like the travelled distance, L R , and the contact area, A C , according to the following equation: NL=Yte,Σ/ACLR(5) 3.2. Surface characteristics of aerosol particles and worn surfaces The collected aerosol particles exhibited a wide size distribution, ranging from approximately 2 μ m to 20 nm, as illustrated in Fig. 6. Larger particles were observed as agglomerates, indicating the aggregation of smaller primary particles. These primary particles displayed an irregular polygonal morphology, a characteristic feature of fractured cementitious materials [44]. Notably, EDS analysis revealed a consistent and uniform elemental composition across significant surface areas exceeding 1000 μ m 2 . Additionally, sporadic occurrences of discrete fiber fragments were identified within the samples, as exemplified in the image of sample A. Backscattered electron (BSE) imaging using scanning electron microscopy revealed the presence of protruding fibers on the worn surfaces. These fibers could be categorized into two distinct types based on their morphology and contrast with the surrounding cementitious matrix. Type I fibers displayed an irregular shape and exhibited negative contrast compared to the matrix (indicated by purple arrows in Column I of Figs. 7 and 8a). This suggests a polymeric composition. In contrast, Type II fibers (highlighted by orange arrows in Figs. 7 and 8b) appeared straight and brighter than the matrix, indicating a higher density compared to the Type I fibers. These characteristics are consistent with the presence of glass fiber reinforcement within the polyurethane fiber components. However, SEM-EDS analysis did not provide definitive evidence regarding the composition of the fibers. The elemental composition of the fibers was found to be similar to that of the surrounding cementitious matrix. This finding is unsurprising given the expected composition of glass fibers. Examination of the worn surfaces using SEM revealed prominent abrasion marks aligned with the sliding direction (Fig. 7, Column I). These marks consisted of alternating brighter and smoother plateaus and darker valleys filled with debris particles in the micrometer range. Both the cementitious matrix and the fibers displayed clear evidence of substantial grinding. However, the two fiber types exhibited distinct wear behaviors. Polymer fibers (Type I, indicated by green arrows) underwent plastic deformation and tearing during abrasion. This resulted in the formation of irregular flake-like debris scattered across the worn surface. In contrast, Type II fibers (highlighted by orange arrows) exhibited brittle fracture behavior, leading to the generation of elongated, sharpedged submicron-sized fragments. Qualitative analysis of the accumulated wear debris on the sides of the wear track (Fig. 7, Column IV) revealed variations in size distribution with changes in PUF aggregate concentration. The significant agglomeration of debris particles precluded a quantitative analysis. Compared to the reference mortar, where debris ranged from tens of nanometers to tens of micrometers, mortars with increasing PUF content displayed a trend towards smaller debris with a more regular, polygonal shape. Interestingly, at a PUF concentration of 60 %, the worn surface exhibited a dramatic change in topography (Fig. 8c). This included highly irregular surfaces with large notches measuring in millimeters along the wear track. Additionally, a network of cracks was observed on finely ground plateaus (Fig. 7, Column II for sample C). This increased deterioration of the mortar can be attributed to the mismatch in mechanical properties between the cement and polymer phases, as well as the potential deficiency of fine aggregate materials within the PUFs themselves. It is important to note that, as recycled waste materials, PUFs may contain various fillers (e.g., BaSO 4 , glass fibers), adhesives, and other constituents, as exemplified by the material shown in Fig. 8d. The presence and influence of these additional components on wear behavior warrant further investigation. Fig. 9 presents the Raman spectra. The spectra of Ref and A samples exhibit a combination of bands corresponding to the mortar matrix. These bands primarily corresponded to the deformation modes of SiO 2 vibration bands [45–47]: Si-O-Si deformation (mode A 1 ) and rotation of the tetrahedra around their twofold axis perpendicular to the c axis (mode E). For samples B and C, additional broad bands were observed. These bands are indicative of stretching and deformation vibration modes of organic groups associated with the PUFs. It is noteworthy that the mortar bands of sample C were barely distinguishable due to a prominent broadband fluorescence emission observed below 1200 cm −1 . Consequently, the spectrum of sample C is not included in Fig. 9. Furthermore, Bruckmoser et al. [48], reported that a significant broadening of vibration bands associated with organic groups can indicate substantial structural degradation in polyurethane resulting from aging, exposure to ultraviolet light, and thermal cycling. Considering the use of recycled polyurethane in this study, such degradation was anticipated. Fig. 5. a) The cumulative emission of particles within the size range of 10–400 nm and the total volume of worn material; b) the specific number of emitted aerosol particles per unit volume of worn material. F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 5 A comparison of Raman spectra between pristine and worn surfaces revealed that the remaining mortar matrix on the worn surfaces did not undergo significant structural or chemical transformations as a result of abrasion. However, the bands corresponding to organic components displayed a tendency to disappear after abrasion, accompanied by an increase in fluorescence signals. This observation was particularly pronounced for samples B and C. This phenomenon can likely be attributed to the combined effects of mechanical shredding and mechanochemical degradation of protruding polymer fibers present on the contact surfaces during abrasion. 3.3. Mortars’ porosity The specific surface area, microand mesopore size distribution, and skeletal density of the mortar samples were assessed using a combination of three techniques: dynamic water adsorption, N₂ adsorption, and He pycnometry. All three methods consistently demonstrated a correlation between the concentration of PUF aggregates in the mortars and an increase in both the specific volume of voids and specific surface area. The results obtained from He pycnometry, a highly reliable technique for measuring skeletal density, are presented in Table 1. The void volume measured for the reference mortar in this study aligns well with Fig. 6. Secondary electron images of captured aerosol particles: a), b) and c) reference sample; d) and e) sample A; f) and g) sample B; and h) and i) sample C. F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 6 findings obtained through previous research using optical methods [49] and mercury intrusion porosimetry [50]. Notably, the porosity exhibited a non-monotonic relationship with the PUF concentration, with the most significant increases observed at 20 % and 60 % PUF content. Table 2 displays the contact angles of water on different mortar samples, which was used to calculate Kelvin’s pore radius. The pore size distribution functions obtained from the H 2 O adsorption isotherms and presented in Fig. 10 demonstrate a general consistency with observations reported in previous studies [51–53]. However, it is noteworthy that the maximum peak of the distribution shifted towards smaller pore radii in our study. The influence of PUF aggregates appears to be most pronounced within micropores and small mesopores, with radii below 3 nm. As the PUF concentration increases, a progressive and nearly two-fold increase in the volume fraction of these smaller pores is observed. In the range of pore radii exceeding 1 nm, the pore size distribution functions obtained from N 2 adsorption isotherms displayed a remarkable similarity to those derived from H 2 O adsorption (observe Figs. 10 and 11). Interestingly, a distinct peak was observed below 1 nm in the N 2 adsorption data, which was absent in the water sorption measurements. Above 1 nm, an increase in the PUF concentration correlates with a greater contribution of microand mesopores. This observation is further supported by the Gurvich’s average pore radius calculated as a function of PUF concentration (Fig. 12), which exhibits a similar trend. The Gurvich’s radius was determined as an approximation for cylindrical pores and given by: rGurvich =2V As(6) where V is the total specific volume of pores determined using He pycnometry method and A s is the specific BET area. In both cases, the average pore radii are within the mesopores size range. Fig. 7. SEM images of worn surfaces and debris. The rows correspond to the Reference, A, B, and C samples. Column I – backscattered electron (BSE) images (20 kV) of pristine and worn surfaces; column II – BSE images of worn surfaces; column III – Large Field Detector images of worn surfaces; column IV – BSE images of debris on the sides of the wear track. The blue arrows in column I show the sliding direction. The purple and orange arrows mark fiber structures. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 7 4. Discussion The analytical methods employed in this study were unable to detect a significant number of micrometer-sized fiber fragments within the wear debris. This suggests that the majority of fibers likely undergo either mechanical degradation and fragmentation into submicron-sized particles, or mechanochemical transformations resulting in submicronsized particles, organic volatile products, and/or a tribofilm. The increased water contact angle observed on the worn surface of the mortar with the highest PUF concentration (sample C) compared to the pristine surface suggests the potential formation of a polymer-based tribofilm. However, conclusive evidence for tribofilm formation on other mortars is lacking, as the contact angle differences between worn and pristine surfaces were not statistically significant. Furthermore, the disappearance of Raman bands associated with PUFs and the emergence of fluorescence signals (Fig. 9) indicate tribochemical degradation of polyurethane fibers. SEM analysis corroborates this, showing a refinement of debris on contact surfaces with increasing PUFs concentration. The presence of ultrafine polymer particles within the emitted aerosol can be inferred from the decrease in the deposition velocity of particles ranging from 10 to 60 nm with increasing PUF concentration as depicted in Fig. 3. This variation can be associated with lower specific gravity of polymeric particles compared to cement hydrates. These observations are significant findings of this study. Based on Eq. (7) from [54], lower specific gravity particles like polymers will exhibit a smaller aerodynamic diameter, D ae , compared to cement hydrate particles of similar size and geometry. This translates to a longer airborne residence time for the polymer particles, potentially contributing to their increased presence in the aerosol. The relationship is defined as follows: Dae,1=Dae,2 ρ 1 ρ 2 √(7) where ρ represents the particle density. Therefore, polyurethane particles with a density of 0.91 g/cm 3 will exhibit a geometric diameter 1.66 times larger than that of the mortar particles with a skeletal density of 2.5 g/cm 3 but possessing the same aerodynamic diameter. As deposition predominantly occurs through Brownian diffusion Fig. 8. SEM images of various structural features: a) protruding fibers of type I on sample A; b) a bundle of fibers of type II on sample B; c) large notches on the surface of sample C; d) non-fiber aggregate from waste. Fig. 9. Raman spectra of pristine and worn surfaces for reference sample (a), sample A (b) and sample B (c). Insets show enlarged portions of the corresponding spectra. F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 8 (both Stocks and Reynolds numbers are <<1), the deposition velocity directly correlates with the diffusion coefficient, which, in turn, is inversely proportional to the geometric diameter, D p , of the particle. [55]: Ddif =kTCc 3 πμ Dp χ (8) where k is Boltzmann’s coefficient, T is the temperature, μ is the air viscosity, and χ is the dynamic shape correction factor. Thus, as the proportion of polymer-to-mortar particles with the same aerodynamic diameter in the aerosol mixture increases, the deposition velocity is expected to decrease. Another factor influencing deposition velocity might be particle shape. Studies suggest that elongated, non-spherical particles, likely released from fiber-loaded mortars, exhibit lower diffusion rates and consequently, slower deposition velocities compared to spherical particles of similar diameter [56,57]. The effect of shape on the deposition velocity is determined by the χ parameter in (8), which can range from 1 to 4 depending on the particle’s aspect ratio and orientation. While the presence of ultrafine polymer particles in the aerosol was not directly confirmed, this analysis suggests the hypothesis aligns with the observations. To rule out frictional heating as a contributor to aerosol emission, the maximum contact temperature was estimated using Jaegers’s model [58]. The results indicated a negligible temperature increase (<50 ◦C above ambient), insufficient for pyrolysis, thermal oxidation, or other thermally driven reactions. Therefore, under these conditions, wear and non-thermally driven tribochemical transformations emerged as the dominant mechanisms for aerosol emission. The observed complex aerosol triboemission dynamics can be attributed to wear behavior influenced by the mortar’s mechanical properties. These align with previous reports on PUF-loaded mortar compressive strength [36,59]. Mortar A displayed higher strength than the reference, while mortars B and C showed reductions of 10 % and 21 %, respectively. Surface hardness (Shore C) also decreased with increasing PUF content. These variations in mechanical properties likely stem from microstructural alterations and increased mesoporosity with higher PUF Table 1 Skeletal density, apparent density, porosity and specific void volume: mean value (standard error). Sample Skeletal density Apparent density Porosity The specific volume of voids per mortar mass g cm −3 g cm −3 % cm 3 g −1 Ref. 2.518 (0.0131) 2.317 7.98 (0.04) 0.0344 (1.72 ×10 −4 ) A 2.511 (0.0144) 2.179 13.2 (0.08) 0.0606 (3.67 ×10 −4 ) B 2.491 (0.0077) 2.142 14.0 (0.04) 0.0654 (1.87 ×10 −4 ) C 2.478 (0.0067) 2.018 18.6 (0.05) 0.0922 (2.48 ×10 −4 ) Table 2 Contact angles of water on mortars samples (mean value ±standard error). Surface Ref A B C pristine 43.9 ±2.2 51.7 ±1.3 31.8 ±5.6 49.5 ±3.6 worn 53.7 ±8.3 67.9 ±15.4 25.7 ±3.1 78.6 ±3.3 Fig. 10. The pore size distribution determined from H 2 O desorption isotherms using Barett-Joyner-Halenda method. Fig. 11. Pore size distribution determined from N 2 desorption isotherms using BJH analysis. Fig. 12. Gurvich’s average pore radius as a function of PUFs concentration determined for H 2 O and N 2 BET adsorption isotherms. F. Husanu et al. Cement and Concrete Research 185 (2024) 107650 9