Characteristics of fume components and dynamic emission behavior of crumb-rubber-modified asphalt

1 Introduction

Incorporating crumb rubber derived from end-of-life tires into asphalt pavements provides an effective pathway for tire-waste valorization while offering potential improvements in pavement performance. However, the production and application of crumb-rubber-modified asphalt (CRMA) typically involve elevated temperatures during rubber incorporation, mixing, swelling, and storage. Under these thermal conditions, volatile constituents of the base asphalt may be released together with low-molecular-weight compounds and sulfur-containing species associated with crumb rubber. The composition and intensity of the resulting fumes vary with processing conditions, creating concerns for emission control and occupational exposure (Tang et al., 2022; Li et al., 2024a). Previous studies have demonstrated that asphalt source and mixing temperature are major determinants of fume composition and emission intensity, while recent reviews have further identified binder composition, heating duration, and sampling/analytical protocols as important sources of variability in asphalt volatile organic compound (VOC) measurements (Mo et al., 2019; Li et al., 2024b).

Recent studies have increasingly moved from laboratory-scale characterization toward emission assessment under production and paving conditions. Plant-scale dry-process studies have coupled gaseous-emission monitoring with functional performance tests, while field investigations have measured gases, particulate matter, and odor during paving with end-of-life tire rubber (Bueno et al., 2021; Chlebnikovas et al., 2023). Life-cycle risk assessment have also been employed to evaluate potential environmental and occupational implications of emissions generated during asphalt production and paving (Zanetti and FARINA, 2022). More recently, high-resolution mass spectrometry has been used to compare recycled mineral, carbonaceous, and geopolymer-based emission-reduction agents, demonstrating that suppression efficiency is strongly dependent on material structure and adsorption characteristics (Borinelli et al., 2024). Collectively, these studies have expanded asphalt fume research from simple binder-level screening toward process-scale emission characterization and control. However, direct comparison among studies remains difficult because of substantial differences in mixture composition, processing scale, sampling environment, and analytical targets. Moreover, field-scale exposure studies have shown that paving temperature, binder content, occupational task, and work rate can significantly affect worker exposure, whereas comparisons between controlled chamber measurement and ambient environments indicate that laboratory-derived concentrations cannot be directly interpreted as field-level air-quality or occupational exposure metrics (Germin-Aizac et al., 2023; Lostier et al., 2025).

In China, recent work has increasingly focused on improving the quantitative methods and mechanistic interpretation of asphalt and CRMA emissions. Advances in sampling procedures, internal standards, selected-ion monitoring, calibration curves, and recovery validation have strengthened GC-MS and HS-GC-MS measurements for PAHs, benzenes, benzothiazoles, and other rubber-associated constituents (Zhang et al., 2024a; Tang et al., 2025). Experiments have further demonstrated that heating history, temperature, residence time, rubber content, rubber particle size, and stabilizer dosage can substantially influence both the magnitude and temporal evolution of emissions (Tang et al., 2022; Li et al., 2024a; Wang et al., 2023). Molecular simulation further suggests that crumb rubber can adsorb light asphalt fractions, whereas excessive swelling or degradation may promote the release of toxic species at high temperature (Xie et al., 2024a). These studies represent an important transition from reporting end-point chemical compositions toward understanding the dynamic and process-dependent nature of CRMA emissions. Related asphalt-scale studies have further shown that mineral composition and surface energy govern asphalt-aggregate adhesion, while castor-based bio-oil can alter asphaltene aggregation behavior (Zhou et al., 2026a; Zhou et al., 2026b). These findings support the combined use of molecular and macroscopic evidence when interpreting asphalt behavior.

Emission-control research in China has likewise progressed from single-component adsorbents toward multifunctional and performance-balanced suppression systems. Nanoclay, attapulgite, warm-mix/organic-montmorillonite blends, and composite agents combining porous adsorption with chemical capture have been investigated for suppressing VOCs, odorants, and H₂S emissions (Liu et al., 2023; Wei et al., 2024; Xie et al., 2024b; Wang et al., 2024). Recent studies have assessed composite deodorizing agents, activated carbon, expanded graphite, and zinc ricinoleate together with rheological and storage performance (Xie et al., 2024c; Lin et al., 2025). The evidence indicates that emission reduction is compound-specific and must be evaluated alongside asphalt engineering properties.

Important limitations remain. Sampling media, collection duration, temperature, gas flow, sample mass, calibration procedures, and reported indicators are not yet standardized, limiting reproducibility and quantitative comparison across studies (Tang et al., 2025; Shi et al., 2025). Most studies still rely on laboratory or discrete sampling, while VOCs, sulfur-containing gases, particulate matter, and odor are often evaluated separately. However, asphalt-fume emission is inherently time dependent rather than a static response that can be adequately represented by a single sampling point. Previous studies have shown that CRMA emissions evolve substantially with processing time and heating history, indicating that time-resolved monitoring is necessary to capture transient emission maxima, subsequent decay, and cumulative emission behavior (Tang et al., 2022; Wang et al., 2023; Gong et al., 2026). Comparisons between base asphalt and CRMA do not always control binder source, and laboratory concentrations should not be directly interpreted as plant emissions or worker exposure. Source attribution, identification of priority pollutants, minimum reporting requirement, and translation from laboratory measurements to field conditions therefore remain insufficiently resolved.

To address these gaps, particularly the limited temporal resolution of conventional discrete sampling, this study developed an integrated experimental framework combining controlled fume generation, real-time monitoring, and offline collection. Experimental repeatability was evaluated under controlled sample mass, temperature, stirring conditions, and thermal history. Dynamic VOC responses were then characterized across four base asphalts and between ZH90# and CRMA prepared from the same binder with 20 wt% crumb rubber. The temperature- and time-dependent responses of H₂S and SO₂ were subsequently quantified for CRMA, and GC-MS was used to compare detectable organic compounds in the two fume samples. The emission behavior was evaluated using complementary indicators, including maximum concentration, time to peak, post-peak decay, late-stage response, and cumulative response, together with differences in chemical composition.

2 Materials and methods

2.1 Raw materials

Four penetration-grade 90 asphalt binders, designated as ZH90#, SK90#, JB90#, and GX90#, were selected for this study. The crumb-rubber-modified asphalt was prepared using ZH90# as the base binder and commercial 40-mesh crumb rubber derived from end-of-life tires, supplied by Shandong Weiwu, China. The crumb rubber was incorporated at 20 wt% of the base-asphalt mass. No additional desulfurization, activation, or other pretreatment was introduced during the laboratory preparation of CRMA. The basic properties provided for the crumb rubber are summarized in Table 1.

The basic physical properties of the asphalt binders are presented in Table 2.

TABLE 1. Basic properties of the crumb rubber used in this study.

TABLE 2. Basic properties of base asphalt and crumb-rubber-modified asphalt.

2.2 Experimental design

An experimental program combining real-time monitoring and offline compositional analysis was used to evaluate the effects of binder source, heating temperature, and crumb-rubber incorporation (Figure 1). The program comprised system-repeatability testing, dynamic monitoring of VOCs and sulfur-containing gases, and GC–MS comparison of ZH90# and its corresponding crumb-rubber-modified asphalt.

System repeatability was evaluated using duplicate ZH90# tests at 120, 140, 160, 180, and 200 °C. VOC concentrations were recorded every 10 s for 1.5 h, and a relative deviation below 20% was considered acceptable. The confirmed system was then applied to ZH90#, SK90#, JB90#, GX90#, and the matched crumb-rubber-modified asphalt at the same five temperatures, giving 25 test conditions with two replicates per condition. Each test used 500 ± 5 g of asphalt; monitoring was conducted after the sample stabilized within ±3 °C of the target temperature while being stirred at 300 rpm. For the modified asphalt, crumb rubber was added at the target temperature and monitoring began after 10 min of mixing, swelling, and pre-purging.

The analysis compared the four base asphalts at each temperature, each material across temperatures, and ZH90# with the modified asphalt prepared from the same binder. VOC responses were summarized by the maximum recorded concentration, time to peak, post-peak evolution, late-stage response, and cumulative response. H₂S and SO₂ were screened for the four base asphalts and then monitored for the modified asphalt over 120 °C–200 °C. GC–MS was performed on ZH90# and the matched modified asphalt under identical generation, sampling, and analytical conditions to compare chromatographic profiles and tentatively identified compounds. Real-time readings and GC–MS responses were used for relative comparison only.

2.3 Dynamic fume monitoring system

2.3.1 Real-time detection module for VOCs and sulfur-containing gases

The real-time module (Figure 2) used a sealed, thermostatically heated separable flask with mechanical stirring. The outlet gas passed through a 0.22 μm hydrophobic PTFE membrane, an ice–water cooling section, and a moisture-removal trap before entering the detector. The gasket and clamp maintained airtightness, while filtration and cooling protected the detector from particles, moisture, and excessive temperature.

2.3.2 Collection system for VOCs and sulfur-containing compounds emitted from crumb-rubber-modified asphalt

Following Wang et al. (Wang et al., 2024), the collection train comprised the common fume-generation unit, a PTFE membrane filter, a silica-gel drying tube, a gas-washing bottle containing 50 mL of cyclohexane, a safety bottle, and an air pump (Figure 3; Figure 4).

2.3.3 Real-time monitoring procedure

To standardize thermal history, the asphalt was heated at 140 °C for 4 h, homogenized, and divided into smaller containers. Before testing, each portion was reheated at 140 °C for 2 h, mixed, and transferred to a separable flask at a mass of 500 ± 5 g.

The flask was heated to the target temperature and assembled with the sealing gasket, stirring unit, thermometer, and outlet line. For modified-asphalt tests, crumb rubber was added gradually and mixed before the vessel was sealed. Once the temperature stabilized within ±3 °C, the sample was stirred at 300 rpm. The modified asphalt was mixed and swollen for 10 min while the system was pre-purged at 500 mL/min; formal monitoring then began after membrane filtration, cooling, and moisture removal.

2.3.4 System repeatability and measurement limitations

For the repeatability assessment, VOC concentrations were recorded every 10 s in duplicate 1.5 h tests. A relative deviation below 20% was used as the acceptance criterion, and the replicate ZH90# curves showed relative deviations below 10% during the principal response stages (Figure 5). Instrument readings are equivalent concentrations determined by the specified flow configuration and calibration and were used only for comparison among laboratory conditions, not as field emission limits or occupational exposure concentrations.

2.4 Fume collection and GC–MS analysis

After 10 min of pre-purging, the drying tube, gas-washing bottle, safety bottle, and pump were connected in series. Fumes were collected in 50 mL of analytical-grade cyclohexane at 250 mL/min for 30 min. The solution was mixed, and a 5 mL aliquot was passed through a 13 mm, 0.22 μm hydrophobic PTFE syringe filter and transferred to a GC–MS vial.

Samples were analyzed with a Shimadzu GCMS-QP2020 equipped with a TG-5MS capillary column (Cao et al., 2021). Helium was used at 1.5 mL/min with a 5:1 split ratio. The oven was held at 40 °C for 3 min, heated to 200 °C at 10 °C/min, and held for 11 min; the total run time was 30 min with a 3 min solvent delay. The mass spectrometer operated in full-scan mode (m/z 35–500) at an ion-source temperature of 230 °C.

Compounds were tentatively identified by matching spectra with the NIST library. Fifteen PAHs were also screened using external standards (Zhang et al., 2024a; Hao and WANG, 2023), with a retention-time difference within ±0.1 min used as the identification criterion. The analysis supported tentative identification and relative response comparison rather than absolute quantification (Figure 6).

2.5 Extraction of dynamic characteristic parameters and statistical analysis

VOC profiles were analyzed over the common 10–2,400 s window at 10 s intervals, whereas H₂S and SO₂ were analyzed over 0–40 min at 5 min intervals. The GC–MS dataset comprised 3,730 points from 3.510 to 22.155 min at 0.005 min intervals; chromatographic retention time was not interpreted as fume-emission time.

For VOCs, the extracted parameters were maximum concentration (C_max),time to peak (t_max),late-stage mean concentration (C_s), and area under the curve (AUC). C_max and t_max were determined within 10–2,400 s, and C_s was the mean of the 61 measurements from 1800 to 2,400 s. For H₂S and SO₂, the maximum recorded value and its occurrence time were obtained over 0–40 min, and the late-stage mean (C_late) was calculated from the 30, 35, and 40 min measurements. The late-stage mean concentration (C_s) and AUC were calculated using Equations 1, 2, respectively:

C s = 1 n ∑ C ᵢ (1)

A U C = ∑ C ᵢ + C ᵢ + 1 2 t ᵢ + 1 ‐ t ᵢ (2)

AUC was calculated by trapezoidal integration and reported as mgs/m³ for VOCs and ppmmin for H₂S and SO₂. It represents cumulative response within the specified window, not emitted mass. A maximum observed at a window boundary was reported only as the maximum recorded value and was not extrapolated.

Temperature and crumb-rubber effects were expressed as absolute difference, relative increase, and ratio; the effect of each 20 °C temperature increment was calculated separately. The absolute difference (Equation 3), relative increase (Equation 4), ratio (Equation 5), and adjacent-temperature effect (Equation 6) were calculated as follows:

Δ X = X R A − X Z H (3)

Relative increase

I X = X R A ‐ X Z H X Z H × 100 % (4)

Ratio

R X = X R A X Z H (5)

Adjacent-temperature effect

G X T = X T ‐ X T ‐ 20 X T ‐ 20 × 100 % (6)

RA denotes the crumb-rubber-modified asphalt, and ZH denotes the corresponding ZH90# base asphalt.

Two replicate tests were conducted for each material–temperature condition. The concentration–time profiles and the derived dynamic parameters were used primarily for descriptive comparison among asphalt types and temperature conditions. Consecutive time points within a single concentration–time profile were not treated as independent replicates.

To eliminate baseline offsets between the two TIC profiles, a centered moving window of 1.0 min was applied. The local baseline, I b t R, was estimated as the fifth percentile of the signal intensities within each window, and the resulting baseline was subsequently smoothed using a 0.10 min moving average. The baseline-corrected signal was defined using Equation 7:

I c t R = max I t R − I b t R , 0 (7)

For the baseline-corrected TIC, the maximum peak intensity and retention time were extracted over 4.00–22.00 min, and the cumulative response (S T I C) was obtained by trapezoidal integration using Equation 8:

S T I C = ∑ I c , i + I c , i + 1 2 t R , i + 1 ‐ t R , i (8)

Baseline-corrected areas were also integrated over 3.51–10, 10–14, and 14–22 min to locate the principal regions of chromatographic difference. These parameters were used only for descriptive comparison, not to quantify compound concentrations or emitted mass.

Individual peak areas in Table 3; Table 4 were generated by the GC–MS software and were used to screen high-response compounds or compare matched compounds. Because no internal standard, calibration curve, or compound-specific response factor was used, peak areas were not summed or interpreted as absolute composition.

TABLE 3. Dynamic VOC emission parameters of ZH90# base asphalt and crumb-rubber-modified asphalt.

C s represents the mean concentration calculated from 61 data points recorded over 1800–2,400 s. AUC was calculated by trapezoidal integration over 10–2,400 s.

For CRMA, at 120 °C, the maximum recorded VOC, concentration occurred at the upper boundary of the 2,400 s analysis window. Therefore, 498.4 mg/m3 represents the maximum recorded concentration within the observation window rather than a confirmed peak concentration, and the true t_max was not resolved and may be ≥ 2,400 s.

TABLE 4. Dynamic H₂S and SO₂ emission parameters of crumb-rubber-modified asphalt.

C late was calculated as the mean of the concentrations recorded at 30, 35, and 40 min. AUC was obtained by integrating the concentration–time profile over 0–40 min. Under all test conditions, t max = 0 min; therefore, C max denotes the maximum recorded value at the onset of formal monitoring within the observation window.

3 Results and discussion

3.1 Temperature-dependent VOCs emission responses of base asphalts from different sources

To distinguish the effects of asphalt source and heating temperature on fume emissions, four base asphalts, designated as JB90#, GX90#, SK90#, and ZH90#, were continuously heated and stirred at 120, 140, 160, 180, and 200 °C. The dynamic VOC emission characteristics of the different asphalt binders were monitored in real time. The resulting VOC emission profiles of the four base asphalts and the crumb-rubber-modified asphalt are presented in Figure 7.

Overall, the VOC responses of the four base asphalts increased with increasing temperature; however, their dynamic profiles did not exhibit an identical two-stage pattern. At certain temperatures, ZH90#, GX90#, and SK90# showed a brief increase or plateau before reaching a maximum, followed by a gradual decline. By contrast, the VOC response of JB90# at 200 °C decreased continuously from a relatively high initial level, indicating that its maximum response may have occurred near the onset of formal data acquisition. After the peak, all curves gradually entered a slowly varying stage, while the late-stage concentrations remained above zero. This behavior suggests the coexistence of rapid volatilization of readily releasable components and slow migration of constituents within the asphalt during sustained heating. Therefore, the temperature effect should not be evaluated solely on the basis of the concentration measured at a single time point. Instead, a comprehensive assessment should consider the maximum recorded value, time to peak, post-peak decay, late-stage concentration level, and overall curve response over the prescribed monitoring period.

From 180 °C to 200 °C, the C_max of ZH90# increased from 395.0 to 672.9 mg/m³, an absolute increase of 277.9 mg/m³ (70.4%), while t_max decreased from 740 to 570 s, C_s increased by 107.5 mg/m³ (36.0%), from 298.5 to 406.0 mg/m³, and the AUC increased by 4.714 × 10⁵ mg· s/m³ (56.6%), from 8.336 × 10⁵ to 1.305 × 10⁶ mg·s/m³. Based on the approximate values in Figure 7, the peak VOC concentrations of JB90# and SK90# increased by approximately 270 mg/m³ (117.4%) and 220 mg/m³ (125.7%), respectively, whereas GX90# increased by only approximately 20 mg/m³ (20.0%). Thus, the VOC response over 180 °C–200 °C was strongly dependent on the source of the base asphalt.

When each base asphalt was compared across the tested temperatures, the VOC response at 200 °C was consistently the highest, and the increase from 180 °C to 200 °C was generally greater than those observed over the lower-temperature intervals. According to the recorded profiles, the maximum VOC concentration of ZH90# was approximately 670 mg/m³ at 200 °C, compared with approximately 390 mg/m³ at 180 °C. For JB90# and SK90#, the maximum concentrations at 200 °C were approximately 500 and 395 mg/m³, respectively, both substantially higher than the corresponding values of approximately 230 and 175 mg/m³ at 180 °C. In contrast, GX90# exhibited a comparatively moderate temperature response, with its maximum concentration increasing only slightly from approximately 100 mg/m³ at 180 °C to approximately 120 mg/m³ at 200 °C. These results indicate that sensitivity to elevated temperatures was strongly dependent on the source of the base asphalt. Therefore, increasing the temperature from 180 °C to 200 °C did not produce an equivalent degree of VOC response enhancement for all base asphalts.

At a given temperature, ZH90# generally exhibited the highest peak concentration and late-stage VOC response, followed by JB90# and SK90#, whereas GX90# showed the lowest response. At 200 °C, the maximum VOC concentration of ZH90# was more than five times that of GX90#. Although the four base asphalts had relatively similar penetration and softening-point values, their VOC responses differed substantially, indicating that conventional physical properties alone are insufficient to directly characterize the fume-emission potential of asphalt binders. These differences are more likely associated with variations in crude-oil source, light-fraction content, the relative proportions of aromatic and resin fractions, molecular-weight distribution, and viscosity at elevated temperatures (Mo et al., 2019; Li et al., 2023a).

Increasing temperature not only elevated the maximum VOC concentration but also altered the post-peak decay behavior and the level of sustained late-stage emissions. ZH90# exhibited pronounced post-peak decay at 160 °C–200 °C, while maintaining a relatively high response at the end of the monitoring period. JB90# showed high initial concentrations at elevated temperatures, together with a prolonged decay period. The response curves of SK90# became progressively and consistently separated as the temperature increased. By contrast, the curves of GX90# obtained at different temperatures were closely spaced, and its overall VOC response remained relatively low. These results indicate that the differences in fume emissions among base asphalts from different sources were reflected not only in their instantaneous emission intensities but also in their cumulative responses over the prescribed monitoring period and their capacity for sustained emissions.

From a physical-process perspective, increasing temperature raises the saturated vapor pressures of light and semivolatile components, while reducing system viscosity and accelerating both internal diffusion and gas–liquid interfacial mass transfer, thereby promoting the transfer of a greater proportion of these components into the gas phase. As the readily volatile components are progressively depleted, the concentration gradient decreases, resulting in a corresponding decline in the instrumental response. The persistence of a measurable late-stage concentration indicates that the slow migration and continued thermal volatilization of higher-molecular-weight components remain ongoing. However, for the nonlinear increase observed at elevated temperatures, the relative contributions of physical volatilization and mild thermochemical transformations cannot be distinguished solely from the overall VOC response profiles. Recent time-resolved and molecular-source studies likewise support the coupled roles of temperature-dependent volatilization, diffusion/desorption, and binder-fraction composition in governing asphalt VOC release (Gong et al., 2026; Jia et al., 2026).

In summary, all four base asphalts exhibited a clear temperature dependence; however, the source of the asphalt also governed the maximum response level, the shape of the dynamic emission profile, and the sustained late-stage response. The VOC responses of ZH90#, JB90#, and SK90# increased particularly markedly over the temperature range of 180 °C–200 °C, whereas GX90# exhibited a comparatively moderate response. These findings indicate that processing-temperature control should be evaluated with consideration of the specific source of the base asphalt, rather than using a single temperature-related increase to generalize the behavior of all asphalt binders.

3.2 Incremental effect of crumb rubber incorporation on dynamic VOC emissions

To evaluate the effect of crumb-rubber incorporation on the dynamic emission behavior of VOCs from asphalt, the VOC concentration–time profiles of the crumb-rubber-modified asphalt and its corresponding ZH90# base asphalt were compared over the temperature range of 120 °C–200 °C. Because the crumb-rubber-modified asphalt was prepared using ZH90# as the base asphalt with the addition of 20 wt% crumb rubber derived from end-of-life tires, differences in VOC emission behavior between the two materials at the same temperature primarily reflect the incremental effect of crumb-rubber incorporation. The dynamic parameters calculated according to the method described in Section 2.5 are presented in Table 3.

As shown in Figure 7D; Figure 8, the VOC concentrations of both ZH90# base asphalt and crumb-rubber-modified asphalt exhibited distinct stagewise variations with heating time. Under most test conditions, the maximum recorded concentration was reached after formal monitoring commenced, followed by a gradual decline and transition into a slowly varying stage. An exception was CRMA at 120 °C, for which the VOC response continued to increase up to the 2,400 s boundary of the analysis window. Therefore, the value reported at 120 °C should be interpreted as the maximum recorded response within the observation window rather than a confirmed emission peak, and the true t_max could not be resolved. These results indicate that the fume response of asphalt is not a constant process but is jointly governed by the rapid release of readily volatile components and their sustained late-stage migration. Compared with ZH90# base asphalt, the crumb-rubber-modified asphalt exhibited higher VOC responses at most temperatures, demonstrating that crumb rubber incorporation altered both the intensity and temporal evolution of the fume response.

As shown in Table 3, the C_max, C_s, and AUC values of the crumb-rubber-modified asphalt were higher than those of the corresponding ZH90# base asphalt at all five temperatures; however, these parameters did not increase strictly monotonically over the range of 120 °C–180 °C. In particular, when the temperature increased from 140 °C to 160 °C, C_max, C_s, and AUC decreased by approximately 8.6%, 18.2%, and 12.6%, respectively, while t_max shifted from 740 to 340 s. This local non-monotonic response may reflect the time-dependent nature of the CRMA emission process. Because formal monitoring commenced after 10 min of mixing, swelling, and pre-purging, a greater proportion of readily volatile or thermally desorbed components may have been released before formal data acquisition at 160 °C. Temperature-dependent crumb-rubber swelling may also have altered the partitioning and subsequent mass transfer of light components among the binder, rubber phase, and gas phase. Therefore, the lower monitored response at 160 °C relative to 140 °C is more appropriately interpreted as a local non-monotonic behavior arising from the combined influence of pre-monitoring release and subsequent mass-transfer processes, rather than as evidence of a simple inverse temperature effect.

At 200 °C, the C_max values of the crumb-rubber-modified asphalt and ZH90# base asphalt were 3,719.1 and 672.9 mg/m³, respectively. Crumb-rubber incorporation therefore increased C_max by 3,046.2 mg/m³ (452.7%), giving a peak ratio of 5.53. The corresponding t_max values were 110 and 570 s. These results show that crumb rubber markedly increased the maximum VOC response and caused the recorded peak to occur earlier within the monitoring window.

In addition to the higher peak concentration, the VOC concentration–time profile of the crumb-rubber-modified asphalt remained consistently above that of the ZH90# base asphalt over the same monitoring period, indicating that crumb-rubber incorporation increased the overall curve response within the prescribed time window. The greater separation between the two profiles at elevated temperatures further suggests that the incremental effect of crumb rubber was reflected not only in an increased recorded peak but also in the sustained persistence of an elevated response over a certain period.

At 200 °C, C_s was 2,284.5 mg/m³ for the crumb-rubber-modified asphalt and 406.0 mg/m³ for ZH90#, an increase of 1878.5 mg/m³ (462.6%) and a ratio of 5.63. Their AUC values were 6.467 × 10⁶ and 1.305 × 10⁶ mg·s/m³, respectively; the modified asphalt therefore showed an AUC increase of 5.162 × 10⁶ mg·s/m³ (395.5%) and an AUC ratio of 4.96. The increases in C_max, C_s, and AUC demonstrate that the effect at 200 °C involved both a higher peak and a sustained elevation throughout the monitoring period.

The VOC concentration of the crumb-rubber-modified asphalt generally remained higher than that of the ZH90# base asphalt after the peak, further suggesting that crumb rubber promoted sustained VOC release during the later stage. Upon heating the base asphalt, light components continuously migrated into the gas phase; as the more volatile fractions were progressively depleted, the emission rate gradually decreased. In the crumb-rubber-modified asphalt, in addition to the volatilization of the base asphalt, the crumb rubber absorbed light asphalt fractions and underwent swelling, thereby altering the viscosity, interfacial state, and mass-transfer pathways of the system. Meanwhile, low-molecular-weight components within the crumb rubber may have gradually diffused outward during sustained heating, allowing the crumb-rubber-modified asphalt to maintain a relatively high VOC emission level even after the post-peak decline.

For the crumb-rubber-modified asphalt, increasing the temperature from 180 °C to 200 °C raised C_max by 2,666.5 mg/m³ (253.3%), from 1,052.6 to 3,719.1 mg/m³. C_s increased by 1,693.3 mg/m³ (286.4%), from 591.2 to 2,284.5 mg/m³, and the AUC increased by 4.742 × 10⁶ mg·s/m³ (274.9%), from 1.725 × 10⁶ to 6.467 × 10⁶ mg·s/m³. The simultaneous increases in all three parameters indicate a pronounced enhancement of the CRMA emission response at 200 °C relative to 180 °C. However, because the temperature interval used in this study was 20 °C, the present data do not resolve the onset temperature or define a specific transition threshold within 180 °C–200 °C. Therefore, this interval should be interpreted only as a high-temperature regime associated with enhanced emissions between the tested temperature points, rather than as a precisely defined sensitive zone.

The enhanced VOC response of the crumb-rubber-modified asphalt at 200 °C may have resulted from the combined effects of several processes. On the one hand, the elevated temperature accelerated the volatilization and thermal desorption of light fractions from the base asphalt and low-molecular-weight additives from the crumb rubber. On the other hand, swelling of the crumb rubber in the asphalt increased its volume and interfacial contact with the surrounding binder, thereby facilitating the migration of volatile substances from the material interior into the gas phase. In addition, elevated temperature may enhance the thermal desorption and release of residual monomers, processing oils, vulcanization-related additives, and other low-molecular-weight components retained within the crumb-rubber matrix. The detection of isoprene, styrene, benzothiophene, benzothiazole, and dibenzothiophene in the subsequent GC–MS analysis confirms that crumb-rubber incorporation altered the composition of the detectable volatile constituents. These compounds may originate from the thermal desorption of pre-existing rubber-bound components and additives. Therefore, their detection by GC–MS should not be interpreted as direct evidence of covalent-bond cleavage or chemical decomposition of the rubber network. Possible structural transformations of the rubber matrix at elevated temperatures cannot be excluded, but further molecular-level characterization would be required to verify their occurrence.

In summary, crumb rubber incorporation increased the peak VOC emission intensity, cumulative response, and sustained late-stage emission level of asphalt, with the most pronounced incremental effect observed at 200 °C. The enhanced VOC emissions from crumb-rubber-modified asphalt were not governed by a single process but were mainly associated with light-fraction volatilization from the base asphalt, enhanced thermal desorption of rubber-bound components, and swelling-induced mass transfer. Possible structural transformations of the rubber matrix cannot be excluded under elevated-temperature conditions, but their occurrence cannot be directly verified based on the present data. In engineering applications, unnecessary exposure of crumb-rubber-modified asphalt to elevated temperatures should therefore be avoided. In particular, prolonged mixing and storage at temperatures approaching 200 °C should be minimized to reduce both the instantaneous VOC emission intensity and the cumulative fume burden.

3.3 Temperature- and time-dependent evolution of sulfur-containing gas emissions from crumb-rubber-modified asphalt

To characterize the emission of sulfur-containing gases following crumb rubber incorporation, the four base asphalts were first subjected to preliminary screening using the same experimental apparatus and detection conditions. Under the experimental and instrumental conditions employed in this study, the H₂S and SO₂ responses of the four base asphalts were close to the lower end of the instrumental detection range. Therefore, the subsequent analysis of sulfur-containing gases focused primarily on the crumb-rubber-modified asphalt. Previous studies have also reported that fumes generated from asphalt at elevated temperatures consist mainly of organic compounds and particulate matter, together with relatively small amounts of sulfur-containing constituents (Autelitano et al., 2017; Sun et al., 2024). Recent CRMA-specific studies have associated H₂S and other sulfur-containing gases with thermal processes involving crumb rubber and have shown that these emissions can be influenced by rubber pretreatment and targeted deodorizing agents (Zhang et al., 2024b; Zhao et al., 2025a; Zhao et al., 2025b). The temperature-dependent H₂S and SO₂ emission profiles of the crumb-rubber-modified asphalt are presented in Figure 9.

The maximum recorded H₂S concentration increased from 10 ppm at 120 °C to 31 ppm at 200 °C, an absolute increase of 21 ppm (210.0%) and a ratio of 3.10. The late-stage mean concentration increased from 5.67 to 21.33 ppm, corresponding to an increase of 276.5% based on the unrounded data. The AUC increased from 265 to 1,035 ppm min, an absolute increase of 770 ppm min (290.6%) and a ratio of 3.91. At all temperatures, the maximum recorded H₂S value occurred at the beginning of formal monitoring and was followed by an overall decline.

For SO₂, C_max increased from 100 ppm at 120 °C to 148 ppm at 200 °C, an increase of 48 ppm (48.0%). Over the same interval, Clate increased by 49.7 ppm (55.8%), from 89.0 to 138.7 ppm, and the AUC increased by 1930 ppm min (51.0%), from 3,785 to 5,715 ppm min. At 200 °C, the SO₂-to-H₂S ratios were 4.77 for C_max, 6.50 for Clate, and 5.52 for AUC, confirming that the SO₂ response was consistently higher than the H₂S response within the monitoring window.

In summary, both H₂S and SO₂ emissions from the crumb-rubber-modified asphalt exhibited temperature-dependent behavior over 120 °C–200 °C, and their concentration–time profiles generally declined after the onset of formal monitoring at all temperatures. At 200 °C, the AUC values of H₂S and SO₂ were 1,035 and 5,715 ppm min, respectively, representing increases of 290.6% and 51.0% relative to those at 120 °C. The corresponding late-stage mean concentrations were 21.3 and 138.7 ppm, respectively. These results indicate that sustained heating near 200 °C enhanced both the initial and cumulative responses of sulfur-containing gases. However, the specific formation pathways require further verification through quantitative analysis of target compounds and characterization of reaction intermediates.

3.4 Chemical composition of asphalt fumes and characteristic crumb-rubber-related compounds

To establish a relationship between the overall response measured by the real-time sensor and the specific chemical composition of the fumes, a comparative analysis was conducted on fumes generated from ZH90# base asphalt and the corresponding crumb-rubber-modified asphalt prepared with ZH90# as the base binder and 20 wt% crumb rubber. The fume-generation temperature, stirring conditions, sampling flow rate, absorption time, and GC–MS operating conditions were kept identical for both materials. As illustrated in Figure 10, fumes generated during the high-temperature mixing of rubber-modified asphalt constitute a multiphase mixture that can be broadly classified into organic and inorganic fractions, including gaseous species, liquid droplets, and solid particles. The organic fraction may contain volatile gases, semivolatile droplets, and condensed or particulate organic constituents, whereas the inorganic fraction may include sulfur-containing gases, water-containing droplets, and particulate matter. In the present study, the fumes were collected using a cyclohexane absorption solution, purified through a 0.22 μm PTFE membrane, and subsequently analyzed by GC–MS. Accordingly, the GC–MS analysis primarily characterized volatile and semivolatile organic compounds that could be enriched by the organic solvent, rather than the complete multiphase composition of the asphalt fumes.

The total ion chromatograms of the fumes emitted from the two asphalt types are shown in Figure 11. Under identical sample-preparation, collection, and instrumental conditions, distinct differences were observed in the chromatographic peak density and retention-time distribution between the two samples. The crumb-rubber-modified asphalt exhibited a greater number of resolved chromatographic peaks within the retention-time range of 4–22 min, indicating increased complexity of the detectable organic components following crumb-rubber incorporation. Because the two TIC profiles exhibited different baseline offsets, their overall response intensities could not be compared directly from their vertical positions in the figure. Therefore, a consistent baseline-correction and integration procedure was required for quantitative comparison.

As shown in Table 5, the maximum baseline-corrected peak intensity decreased from 77,305 counts for ZH90# to 69,166 counts for the crumb-rubber-modified asphalt, a decrease of 8,139 counts (10.5%). In contrast, the cumulative baseline-corrected TIC response increased from 322,420 to 380,323 countsmin, an increase of 57,903 countsmin (18.0%). The baseline-corrected area over 10–14 min increased from 58,639 to 82,401 countsmin, an increase of 23,762 countsmin (40.5%). These results indicate that crumb-rubber incorporation enhanced the distributed chromatographic response across multiple peaks rather than increasing the single highest-intensity peak.

TABLE 5. Baseline-corrected TIC parameters of ZH90# base asphalt and crumb-rubber-modified asphalt.

S TIC was calculated by integrating the complete chromatographic record over the retention-time range of 3.510–22.155 min. The maximum baseline-corrected peak intensity was identified within 4.00–22.00 min. All parameters were obtained using the same local baseline-correction algorithm and were used exclusively for relative comparisons between the two samples.

Table 6; Table 7, list the compounds with relatively large peak areas detected in the fumes from the ZH90# base asphalt and the crumb-rubber-modified asphalt, respectively. As shown in Table 6, the major components detected in the fumes from the ZH90# base asphalt mainly comprised aliphatic hydrocarbons, monocyclic aromatic hydrocarbons, and fused-ring aromatic compounds. Heptane, octane, decane, 1-hexadecene, and eicosane reflect the thermal volatilization of petroleum-derived hydrocarbons with different carbon numbers, whereas styrene, n-propylbenzene, naphthalene, 2-methylnaphthalene, acenaphthene, fluorene, phenanthrene, and fluoranthene indicate that the aromatic fraction exhibited a broad range of volatilities and retention times. These results suggest that the fumes from the base asphalt were dominated by petroleum-derived hydrocarbons, with the detectable composition extending from low-molecular-weight aliphatic hydrocarbons to moderately volatile polycyclic aromatic compounds.

TABLE 6. Major compounds detected in fumes from ZH90# base asphalt.

TABLE 7. Major compounds detected in fumes from crumb-rubber-modified asphalt.

As shown in Table 4, compared with the corresponding base asphalt derived from the same source, the crumb-rubber-modified asphalt fumes not only retained aromatic compounds such as styrene, butylbenzene, cumene, and indene, but also contained isoprene, an isoprene dimer, carbon disulfide, benzothiophene, benzothiazole, methylbenzothiazole, dibenzothiophene, and tetramethylthiuram disulfide. Isoprene and its dimer are closely associated with the polyisoprene structure present in tire rubber and may therefore serve as potential indicator compounds for the thermal release of crumb rubber. Benzothiazole compounds are commonly associated with accelerator systems used in rubber vulcanization, whereas tetramethylthiuram disulfide is a typical sulfur-containing rubber additive. The simultaneous detection of these compounds indicates that crumb rubber incorporation not only altered the relative responses of petroleum-derived hydrocarbons but also introduced additional fume signals associated with rubber polymers, processing additives, and sulfur-containing structures.

Styrene was detected in both fume samples and therefore cannot be used independently as a characteristic marker of crumb-rubber origin. In comparison, the combined occurrence of isoprene-related compounds, benzothiazole derivatives, and multiple sulfur-containing signals, including carbon disulfide, benzothiophene, and dibenzothiophene, provides a more representative indication of the compositional increment associated with crumb rubber incorporation. It should be noted that these compounds are not exclusively derived from crumb rubber. Their diagnostic significance arises from their simultaneous occurrence and changes in relative response, rather than from the presence or absence of any single compound. This interpretation is consistent with recent emission- and odor-tracing studies showing that source attribution is more robust when multicomponent chemical fingerprints are considered rather than relying on a single VOC marker (Li et al., 2025; Li et al., 2023b; Deller et al., 2024).

In terms of retention-time distribution, isoprene, carbon disulfide, isoprene dimer, and several aromatic hydrocarbons were detected in the crumb-rubber-modified asphalt within 4–10 min, whereas sulfur-related compounds, including benzothiophene, benzothiazole, methylbenzothiazole, dibenzothiophene, and tetramethylthiuram disulfide, were detected within 14–22 min. This distribution indicates that the detectable signals from the crumb-rubber-modified asphalt spanned a broad chromatographic retention-time range and exhibited incremental responses across multiple intervals. It should be emphasized that retention time is jointly governed by the properties of the chromatographic column, the oven temperature program, and interactions between the compounds and the stationary phase. Therefore, retention time cannot be used to infer the chronological order in which individual components were released during fume generation, nor can it be directly associated with the early or late stages of the real-time VOC concentration–time profiles.

Overall, the GC–MS results showed that the organic composition of the crumb-rubber-modified asphalt fumes was more complex than that of the corresponding base asphalt from the same source. The influence of crumb rubber incorporation was reflected both in the enhanced responses of several compounds common to the two samples and in the appearance of additional characteristic compounds. Isoprene-related compounds, benzothiazole derivatives, and sulfur-containing heterocyclic compounds constituted a representative combination of crumb-rubber-related constituents. This compositional evidence is consistent with the incremental VOC and sulfur-containing gas responses observed by real-time monitoring, supporting the conclusion that fumes from crumb-rubber-modified asphalt comprise both background emissions from the base asphalt and additional emissions associated with the crumb rubber.

3.5 Dynamic emission mechanisms and engineering implications

Based on the results presented in Sections 3.1–3.4, the dynamic fume response of crumb-rubber-modified asphalt can be primarily interpreted in terms of light-fraction volatilization and diffusion from the base asphalt, enhanced thermal desorption of rubber-bound low-molecular-weight components, and swelling-induced mass transfer. The first process contributes to the background asphalt response, whereas the latter two provide plausible explanations for the incremental response following crumb-rubber incorporation. Possible structural transformations of the rubber matrix at elevated temperatures cannot be excluded; however, the present GC–MS and sensor data do not provide direct molecular-level evidence of their occurrence, and this mechanism therefore remains an alternative possibility requiring further verification.

Formal monitoring commenced after the crumb rubber had been added and subjected to 10 min of mixing, swelling, and pre-purging. Therefore, the recorded starting point represents the sustained emission stage at the prescribed temperature rather than the absolute beginning of the entire crumb-rubber-modified asphalt preparation process. Under most test conditions, the VOC concentration initially increased rapidly and reached a recorded maximum, indicating that continuous heating and stirring temporarily caused light components to enter the gas phase at a rate greater than that at which the fumes were removed through the gas-flow system. For the crumb-rubber-modified asphalt, the swollen interfacial regions formed through the absorption of aromatic fractions by the crumb rubber, together with the desorption of processing oils, plasticizers, and other low-molecular-weight additives, contributed to the elevated early-stage response.

After the recorded maximum was reached, the available pool of readily volatile components was progressively depleted, and the concentration gradient between the material interior and surface decreased, resulting in a gradual decline in the fume response. The increase in viscosity induced by crumb rubber swelling, together with the absorption of certain light fractions by the crumb rubber, may also have restricted subsequent diffusion. The persistence of a non-zero response during the late monitoring stage indicates that slow internal diffusion and sustained desorption of additives remained active, while possible structural transformations of the rubber matrix cannot be excluded as an additional mechanism. The maximum recorded value, cumulative response over the prescribed monitoring period, and late-stage response level respectively characterize the instantaneous emission intensity, emission persistence, and late-stage emission capacity. These parameters should therefore be evaluated jointly when characterizing fume emissions from crumb-rubber-modified asphalt.

Temperature was a key factor governing the rates and relative contributions of the different emission processes, although the VOC response did not increase monotonically at every adjacent temperature. In particular, the crumb-rubber-modified asphalt exhibited lower C_max, late-stage concentration, and AUC at 160 °C than at 140 °C. This local non-monotonic behavior may reflect the combined influence of release occurring before formal monitoring, temperature-dependent crumb-rubber swelling, and subsequent partitioning and mass transfer of volatile components. Within the range of 180 °C–200 °C, however, the increases in the VOC peak concentration, overall concentration–time profile, and late-stage response of the crumb-rubber-modified asphalt became substantially more pronounced. At 200 °C, the peak VOC concentration was approximately five times that of the corresponding ZH90# base asphalt. Meanwhile, the maximum recorded H₂S concentration increased from 10 ppm at 120 °C to approximately 31 ppm at 200 °C, and SO₂ also exhibited its highest response at 200 °C. Given the 20 °C temperature intervals employed in the present study, 180 °C–200 °C should be interpreted only as a high-temperature regime associated with an enhanced emission response, rather than as a precisely resolved transition or sensitive zone (Figure 12).

The GC–MS results provided compositional evidence for the incremental emissions associated with crumb rubber. Isoprene and its dimer are associated with the polyisoprene-based rubber matrix, whereas benzothiazole derivatives and tetramethylthiuram disulfide are related to vulcanization accelerator systems and rubber additives. The release of these compounds can therefore be reasonably explained by enhanced thermal desorption of residual monomers, additives, and other low-molecular-weight components retained within the crumb rubber. The sulfur-containing compounds detected by GC–MS, together with the temperature-dependent H₂S and SO₂ responses, indicate that sulfur-containing constituents contribute to the fume composition. However, these observations do not provide direct evidence of desulfurization or covalent-bond cleavage of the rubber network. Possible structural transformations at elevated temperatures cannot be excluded, although their occurrence requires further verification using direct molecular-level characterization.

Production temperature and high-temperature residence time are critical process parameters for controlling fume emissions from crumb-rubber-modified asphalt. Provided that the requirements for crumb rubber dispersion, swelling, and pavement performance are satisfied, the lowest temperature capable of achieving the intended processing objectives should be adopted, and the durations of high-temperature storage, recirculating heating, and repeated reheating should be minimized. At temperatures approaching 200 °C, particular attention should be paid to preventing localized overheating, and multipoint temperature measurements should be used to identify temperature deviations within the processing system. Process evaluation should consider the maximum recorded value, cumulative response, and late-stage response level simultaneously, rather than assessing emission-reduction effectiveness solely on the basis of a reading at a single time point or changes in odor. Recent fume-control studies similarly show that the effectiveness of suppression materials depends on the targeted chemical class and that emission reduction should be assessed together with rheological and storage performance (Lv et al., 2024; Li et al., 2024c). The dynamic parameters in Tables 3 and 4 further provide a basis for temperature-specific operational control. At the tested temperatures of 120 °C–160 °C, the VOC responses were substantially lower than those observed at 200 °C, although appreciable and sustained emissions were still present. In particular, the unresolved VOC maximum at 120 °C indicates that operation at a lower temperature does not necessarily eliminate prolonged fume release; therefore, ventilation and fume capture should remain active throughout both heating and holding. At 180 °C, the maximum recorded VOC response occurred rapidly (t_max = 40 s), while the maximum recorded H₂S and SO₂ responses occurred at the beginning of formal monitoring, indicating the importance of activating source capture before or at the onset of crumb-rubber processing. At 200 °C, CRMA exhibited the highest C_max, late-stage concentration, and cumulative VOC response, together with the highest cumulative H₂S and SO₂ responses. Accordingly, residence time, repeated reheating, and localized overheating near 200 °C should be minimized, and continuous local exhaust ventilation should be maintained. These operational implications are derived from the relative emission responses obtained under the present laboratory conditions and are not intended to represent regulatory exposure limits.

The fume-capture system should cover the entire processing sequence, including crumb rubber addition, rapid mixing, swelling, transfer, and storage. Local exhaust ventilation should be activated before crumb rubber addition, and stable negative pressure should be maintained during periods of elevated emission response. Enclosed feeding and mixing operations, together with reduced worker residence time near emission sources, can minimize fugitive dispersion. The collected fumes should be treated in stages according to their physical phases and chemical composition. Particulate matter and condensable droplets should first be removed by filtration or condensation, after which suitable VOC adsorption, condensation, or oxidation processes should be selected according to the pollutant concentrations and compositions. Dedicated treatment units for H₂S and SO₂ should also be incorporated. On-site exposure assessment should use continuous or high-time-resolution monitoring that covers the material-feeding and initial mixing stages, thereby avoiding the underestimation of short-duration high-response events that may result from single-point sampling during steady-state production.

The real-time concentrations obtained in this study represent equivalent responses under the specific reactor scale, gas-flow rate, and instrumental calibration conditions employed. These measurements are suitable for relative comparisons among samples and temperature conditions but should not be directly interpreted as emission concentrations at asphalt plants or as worker exposure levels. Laboratory binder tests differ from field-scale production in terms of processing scale, adsorption by mineral aggregates, ventilation conditions, and thermal history. Future studies should quantify key organic compounds using authentic standards and internal-standard methods and should integrate dynamic gas-phase monitoring with measurements of particulate-phase and condensable components, as well as personal exposure sampling. Simultaneous validation should also be conducted at asphalt plants and paving sites. Accordingly, minimizing unnecessary high-temperature exposure and residence time, improving the full-process coverage of fume-capture systems, and implementing component-specific treatment strategies should be prioritized in subsequent field-scale investigations. This limitation is also supported by occupational field measurements and chamber-to-city modeling, both of which show that exposure and ambient impacts depend strongly on paving conditions, work environment, and atmospheric processes (Germin-Aizac et al., 2023; Lostier et al., 2025).

4 Conclusion

Considering the sustained heating experienced by asphalt during practical processing, the dynamic concentrations of VOCs and sulfur-containing compounds emitted from base asphalt and crumb-rubber-modified asphalt were monitored at different temperatures. The resulting fume emission behaviors were analyzed, leading to the following main conclusions:

The replicate concentration–time profiles of ZH90# base asphalt exhibited similar positions of the maximum recorded values and consistent decay trends, with relative deviations below 10% during the principal response stages. These results demonstrate that the repeatability of the experimental system was sufficient for comparisons among the test conditions employed in this study. The instrument readings represent equivalent concentrations under the specific gas-flow configuration, flow rate, and calibration conditions and should not be interpreted as field-scale emission concentrations.

The VOC responses of the four base asphalts generally increased with increasing temperature, whereas the peak magnitude, profile shape, and late-stage response depended on the asphalt source. Under most test conditions, the VOC response gradually decreased after reaching the recorded maximum. In contrast, JB90# at 200 °C exhibited a continuous decline from a relatively high initial level. ZH90#, JB90#, and SK90# showed relatively pronounced increases in response over the range of 180 °C–200 °C, whereas GX90# exhibited a comparatively moderate variation.

Crumb rubber incorporation increased the peak VOC concentration, cumulative response over the prescribed monitoring period, and late-stage response. At 200 °C, the peak VOC concentration of the crumb-rubber-modified asphalt reached 3,719.1 mg/m³, approximately 5.5 times that of the corresponding ZH90# base asphalt. The maximum recorded H₂S concentration increased from 10 ppm at 120 °C to 31 ppm at 200 °C, while SO₂ also maintained a relatively high response under elevated-temperature conditions.

The GC–MS results showed that the organic composition of the crumb-rubber-modified asphalt fumes was more complex than that of the base-asphalt fumes, with characteristic signals associated with isoprene-related compounds, aromatic compounds, and sulfur-containing heterocyclic compounds. The enhanced high-temperature fume response of the crumb-rubber-modified asphalt was primarily associated with light-fraction volatilization, enhanced thermal desorption of rubber-bound components, and swelling-induced enhancement of mass transfer. Possible structural transformations of the rubber matrix at elevated temperatures cannot be excluded, although direct evidence for such structural transformations was not obtained in the present study. From an engineering perspective, unnecessary exposure to elevated temperatures and prolonged storage should be avoided, and appropriate control measures should be developed by integrating local fume capture with composition-specific risk assessment.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Ramadhansyah Putra Jaya, Universiti Malaysia Pahang, Malaysia

Reviewed by: Jue Li, Changsha University of Science and Technology, China

Derun Zhang, Huazhong University of Science and Technology, China