Rejuvenation mechanisms in bituminous RAP mastics: insights from FTIR spectroscopy and multivariate discriminant analysis

Rejuvenation mechanisms in bituminous RAP mastics: insights from FTIR spectroscopy and multivariate discriminant analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Rejuvenation mechanisms in bituminous RAP mastics: insights from FTIR spectroscopy and multivariate discriminant analysis Mohsen Motevalizadeh, Konrad Mollenhauer This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6342505/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study examines the rejuvenation mechanisms in bituminous RAP mastics using Fourier transform infrared (FTIR) spectroscopy and multivariate discriminant analysis. The chemical interactions between RAP binders, virgin bitumen, warm mix additives, and two commercial recycling agents (RJs) were evaluated to determine their effectiveness in restoring aged binders. A hybrid approach combining partial least squares regression and linear discriminant analysis (PLSR-LDA) was applied to extract latent variables, classify samples, and identify critical wavenumbers associated with rejuvenation. The findings indicate that Sylvaroad mitigates oxidation effects, particularly around ~ 1758–1715 cm − 1 , while Storflux primarily influences methyl bending near 1370–1360 cm − 1 . Despite these effects, irreversible oxidative aging remains evident in the 1300–1160 cm − 1 range, linked to oxygen-containing compounds, suggesting that complete restoration of binder properties is unattainable. Additionally, mineral composition markers, particularly in the 489–446 cm − 1 and 590–555 cm − 1 ranges, persist as key indicators for distinguishing varying RAP contents after rejuvenation. Civil Engineering FTIR Spectroscopy Multivariate Discriminant Analysis Machine Learning Classification Partial Least-Square Regression Linear Discriminant Analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 1 Introduction The increasing focus on sustainability and circularity in asphalt production and road construction has driven significant interest in incorporating higher proportions of reclaimed asphalt pavement (RAP) into new asphalt mixtures. RAP utilization presents an opportunity to reduce resource consumption and environmental impact, but it also introduces challenges due to the altered properties of the aged binder coating RAP (referred to as RAP binder). Aging mechanisms, such as oxidation and the loss of volatile components, lead to embrittlement and hardening of the binder, significantly impacting the performance of asphalt mixtures. Without adequate measures to address these changes, the incorporation of RAP can result in premature distresses such as cracking and fatigue cracking, thereby shortening the service life of asphalt structures [ 1 ]. The aging-induced changes in bitumen structure are commonly classified into reversible and irreversible categories. Oxidation, a major irreversible process, results in the formation of oxygen-containing functional groups such as sulfoxide (S = O) and carbonyl (C = O), as well as an increase in large molecular species due to dehydrogenation [ 2 , 3 ]. While reversible changes can be counteracted through rejuvenation, irreversible changes remain permanent, even after applying recycling agents (RAs). These agents are designed to restore the colloidal balance of bitumen by rebalancing the asphaltene-to-maltene ratio [ 4 ], transitioning the binder structure from a “gel-type” to a “sol-gel type” system [ 5 ]. However, excessive use of RAs can create a "sol-type" structure, making the binder more prone to deformation at high temperatures [ 6 ]. RAs introduce distinctive chemical components to restore bitumen's microstructure and rheological properties. Despite its effectiveness, complete restoration of aged binders’ original characteristics is unattainable due to irreversible chemical changes. Therefore, optimizing RA content based on the RAP binders’ aging level is critical to achieving the best restoration level without compromising durability. Several studies have evaluated rejuvenation efficiency using various experimental approaches, including mechanical, rheological, and microstructural assessments. Mechanical testing has demonstrated that RAs enhance the cracking resistance of RAP-containing asphalt mixtures while maintaining rutting resistance and moisture susceptibility [ 7 ]. Measures such as the penetration index have been proposed as indicators of RA effectiveness [ 8 ]. Rheological evaluations, involving dynamic shear rheometer (DSR) and bending beam rheometer (BBR) tests, have been used to assess the restoration of binder properties at both high and low temperatures [ 9 , 10 ]. These studies have highlighted the ability of RAs to reduce the performance grade of aged binders in recycled mixtures. Microscopic techniques, such as atomic force microscopy (AFM), have investigated the morphological and microstructural changes in aged binders induced by RAs [ 11 – 13 ]. Significant differences have been reported in the effectiveness of various RAs, including vegetable oil, aromatic extracts, and bio-based agents, in altering the microstructure of aged binders [ 12 , 14 , 15 ]. Similarly, Fourier transform infrared (FTIR) spectroscopy and SARA fractionation have been utilized to characterize the chemical changes associated with rejuvenation. FTIR spectroscopy, in particular, has been used to monitor the absorption intensities of functional groups before and after rejuvenation, providing insights into the oxidative aging processes and the effectiveness of RAs [ 16 ]. In this context, various methods have been proposed including the calculation of band areas as described in [ 17 ], employing an aging index tied on the carbonyl, sulfoxide, and aliphatic band [ 18 ], and measuring the carbonyl and sulfoxide indices per [ 19 ]. However, questions remain regarding the reliability of FTIR indices for distinguishing between different RAs and their effectiveness due to their dependence on functional groups reflecting oxidative aging [ 20 ]. Given that the migration of aged components into the blended binder appears as chemical alterations, FTIR spectroscopy can be a suitable method for uncovering the mechanisms underlying the rejuvenation process. Using the extensive information provided by the FTIR spectroscopy, substantial information can be gained to characterize the changes induced by RJ incorporation and the blending of RAP binder with virgin bitumen. Beyond the commonly studied features and wavenumbers, such as carbonyl and sulfoxide function groups, chemometric analyses employing data science techniques—specifically multivariate discriminant analysis—have attracted attention for their potential to identify less prominent features (e.g., wavenumbers with lower absorption intensities) that play critical roles [ 21 ]. This approach offers several advantages, including differentiating bitumen samples based on type and aging state [ 21 ], identifying bitumen sources and refining source [ 22 ], detecting polycyclic aromatic hydrocarbons (PAHs) in bitumen samples [ 23 ], assessing aging evolution rates in bitumen mastics [ 2 ], and differentiating bitumen mastics based on the types of incorporated mineral fillers [ 24 ]. Overall, this analytical approach enables more accurate differentiation of bituminous samples according to specific criteria and provides insights into the underlying mechanisms driving these differences through feature engineering methodologies. This can be beneficial in recognizing the chemical changes induced in bitumen structure from a chemical point of view. Multivariate discriminant analysis is a statistical approach that integrates data mining and machine learning techniques to classify and distinguish bituminous samples with predefined labels based on their chemical characteristics. This method identifies underlying patterns and relationships among wavenumbers that influence differentiation, considering variables such as aging level, RAP content, and additive type. The fundamental principle of multivariate discriminant analysis involves projecting data onto discriminant functions to enhance separation between groups. This process typically consists of two main steps: reducing the dataset’s dimensionality and classifying the reduced dataset according to predefined classes. By combining distinct algorithms, this hybrid technique aims to maximize the separation between predefined classes, facilitating dataset analysis and feature extraction. Various algorithms are employed for dimensionality reduction, including factor analysis (FA), principal component analysis (PCA), and partial least square regression (PLSR). The underlying principles and mathematical formulations of these algorithms are elaborated in [ 3 ]. All dimensionality reduction methods focus on generating new features–latent variables (LVs) or principal components (PCs)–to simplify complex spectral data while retaining most of the variance. PCA identifies directions through PCs that capture the maximum variability in the dataset without considering labels, aiming solely to maximize variance. FA, on the other hand, seeks LVs that explain relationships within the dataset. Unlike these two methods, PLSR incorporates both predictors and responses, maximizing covariance between them to identify LVs that strengthen the relationship between independent and dependent variables. The primary objective of PLSR is to pinpoint features that contribute most to class separability. The effectiveness of these dimensionality reduction techniques in uncovering patterns within FTIR spectroscopy data of bituminous samples has been explored, proposing that the PLSR algorithm can show promises in generating LVs while providing variance importance in projection (VIP) scores, offering an intuitive approach to feature selection [ 2 ]. Following successful dimensionality reduction, linear discriminant analysis (LDA) has emerged as a robust classifier due to its inherent ability to generate characteristic components that project distinct classes onto a Cartesian coordination system. In the context of evaluating the effectiveness of rejuvenation, the mechanisms governing the performance of RJs remain insufficiently understood despite promising outcomes reported through experimental and mechanical investigations. To study the interactions between RJs and blended binders, this study aims to address these gaps by focusing on two critical aspects: (i) the extent to which the RAP binder fraction interacts with virgin bitumen to influence the properties of the blended binder, and (ii) the mechanisms by which distinct RJs alleviate the effects of aged components transferred from the RAP binder into the blended binder. The blending and rejuvenation phenomena in recycled asphalt mixtures will be examined using FTIR spectroscopy to address the research objectives and bridge existing knowledge gaps. The resulting data will be analyzed through multivariate discriminant analysis, specifically by employing PLSR coupled with LDA (PLSR-LDA), to uncover underlying patterns and principal mechanisms governing the interactions between RJs and aged components within RAP binders. The specific goals of this research are outlined as follows: 1- Investigating whether the blended binder in recycled mixtures contains aged and oxidized compounds in proportion to the recycling rate. 2- Determining the extent to which RJs can mitigate the presence of aged components and elucidating the mechanisms involved. 3- Identifying the types of aged components that can be retrieved and retained after rejuvenation. 2 Materials and methods The research aimed to address key objectives by preparing experimental samples at bitumen mastic scale. Fine RAP fractions, passing a 90 µm sieve size, were blended with basalt mineral fillers, virgin bitumen, two warm mix additives, and two commercial RJs. A constant filler-to-binder volumetric ratio of 27% was maintained to ensure consistent findings, in line with recommendations for dense-graded bituminous mixtures featuring a nominal maximum aggregate size of 19 mm [ 25 ]. Specific gravity measurements were conducted for virgin filler, RAP filler, recovered RAP filler, and Asphamin filler, the latter being characterized by its filler form and moisture content. The base bitumen had a penetration grade of 50/70. Sasobit and Asphamin were incorporated at dosages of 3% and 0.3% by weight, respectively. RAP contents of 0%, 30%, 50%, 70%, and 100% were employed to examine a wide inclusion range. The RAP filler, sourced locally, had a binder content of 12.6%, determined using the ignition method per ASTM D6307 specifications. Each mastic formulation was back-calculated based on the measured specific gravity of the components to achieve the target filler-to-binder volumetric ratio of 27%. Table 1 outlines the 14 prepared mastic compositions, detailing the volumetric concentrations of RAP filler, virgin bitumen, RAP binder, virgin filler, and Asphamin, alongside the filler-to-binder volume ratios. The selected volumetric ratio led to mass ratios ranging from 1.1:1 to 1.18:1, aligning with dense-graded asphalt concrete compositions. Mastics were categorized based on their components: those incorporating only virgin fillers and/or warm mix additives were labeled "VM". In contrast, mastics containing RAP fillers represented varying levels of RAP replacement. For complete virgin filler replacement, virgin bitumen was used to maintain the target volumetric filler-to-binder ratio. The mastics were prepared using low-shear mixing at 160°C for unmodified samples and 135°C for those containing warm mix additives. Virgin filler was preheated at 160°C, while RAP fines were conditioned at 110°C for 2 hours [ 25 ]. After component mixing, mechanical stirring was performed at 900 rpm. To prevent phase separation between bitumen and filler, the blends were poured into 4 mm thick metal molds. Furthermore, two commercial rejuvenators (e.g., Storflux and Sylvaroad) were utilized. The optimal dosages for each were determined as outlined in Section 3 through a binder-fast-characterization-test (BTSV), according to European Standard EN 17643, conducted using a dynamic shear rheometer (DSR). Based on the determined optimal contents and the RAP binder amounts specified for each mastic variant in Table 1 , the required rejuvenator quantities were incorporated directly during the mastic preparation process. Table 1 Composition of Mastics Mastic ID Mastic characteristics Rejuvenation RAP filler volumetric concentration Volumetric percentages in bituminous mastic Filler-to-binder volumetric ratio Virgin binder RAP binder Virgin filler Recovered RAP filler Asphamin VM Virgin mastic Unrejuvenated; Storflux and Sylvaroad RJs 0 72.91 0.00 27.09 0.00 0 0.27 Saso-VM Virgin mastic with Sasobit 0 72.91 0.00 27.09 0.00 0 Asph-VM Virgin mastic with Asphamin 0 72.19 0.00 22.43 0.00 5.38 RAP-30 Mastic with 30% RAP 30 69.16 3.75 18.97 8.13 0 RAP-50 Mastic with 50% RAP 50 66.66 6.25 13.55 13.55 0 RAP-70 Mastic with 70% RAP 70 64.16 8.75 8.13 18.97 0 RAP-100 Mastic with 100% RAP 100 60.40 12.50 0.00 27.09 0 Saso-RAP-30 Mastic with 30% RAP and Sasobit 30 69.16 3.75 18.97 8.13 0 Saso-RAP-50 Mastic with 50% RAP and Sasobit 50 66.66 6.25 13.55 13.55 0 Saso-RAP-70 Mastic with 70% RAP and Sasobit 70 64.16 8.75 8.13 18.97 0 Saso-RAP-100 Mastic with 100% RAP and Sasobit 100 60.40 12.50 0.00 27.09 0 Asph-RAP-30 Mastic with 30% RAP and Asphamin 30 68.34 3.85 14.07 8.34 5.40 Asph-RAP-50 Mastic with 50% RAP and Asphamin 50 65.77 6.42 8.51 13.91 5.40 Asph-RAP-70 Mastic with 70% RAP and Asphamin 70 63.20 8.98 2.95 19.47 5.40 Asph-RAP-100 Mastic with 100% RAP and Asphamin 100 61.85 10.33 0.00 22.40 5.42 The mastics underwent chemical measurement using FTIR spectroscopy with a Bruker Alpha device equipped with an attenuated total reflection (ATR) unit. Spectral intensities were recorded over wavenumbers ranging from 4000 to 400 cm − 1 with 24 scans for each spectrum. To ensure reproducibility, five samples of each testing variant were prepared, each measured five times, according to [ 23 ]. 3 Optimal RJ content To determine the appropriate dosage of rejuvenators for restoring the desired properties of virgin bitumen, BTSV testing was utilized to assess the rheological characteristics of bitumen samples according to [ 26 ]. This method involves measuring bitumen's complex modulus and phase angle over a continuous temperature increase. Key parameters derived from this test include: (i) equi-shear modulus temperature at which the complex modulus equals 15 kPa (T BTSV ) and (ii) the corresponding phase angle (δ BTSV ). According to methodologies outlined in [ 27 , 28 ], the examined bitumen samples can be differentiated into classes reflecting their penetration grades and modification level. This investigation performed BTSV testing on virgin bitumen (50/70 penetration grade) and binder recovered from RAP, as illustrated in Fig. 1 a,b. The results showed that the virgin binder showed a T BTSV of 51.53 ℃ and a δ BTSV of 78.89°. On the other hand, the RAP binder showed a T BTSV of 68.61 ℃ and a δ BTSV of 76.91°, displaying the effect of aging, including increased stiffness and embrittlement. These findings, mapped according to the classification framework detailed in [ 27 , 28 ], are presented in Fig. 2 . The results of the BTSV testing indicated that the binder characteristics were positioned, with the combination of T BTSV and δ BTSV correctly reflecting the penetration grade of the bitumen. In contrast, the RAP binder characteristics, represented by the red dot, showed a significant increase in binder stiffness due to aging, positioning the sample closer to the category of bitumen with a 20/30 penetration grade. To counteract the aging effects, various contents of Sylvaroad and Storflux were blended with RAP binder, and the resultant blends were subjected to BTSV testing. The projections from the BTSV testing revealed that employing 10% w/w Sylvaroad and 15% w/w Storflux significantly restored the desired characteristics of the RAP binder, although failed to return to the 50/70 penetration grade category fully. The data in Fig. 2 show that increasing the RJ contents reduced the T BTSV while only slightly restored δ BTSV , thereby improving the penetration grades. Based on the changes depicted in Fig. 2 , the optimal RJ contents were determined for the preparation of rejuvenated RAP mastics, as outlined in Table 1 . The testing results for the RAP binder with 15% w/w Storflux and 10% w/w Sylvaroad are presented in Fig. 1 c,d. 4 Exploration of chemical markers 4-1- Formal analyses Figure 3 a-c illustrates a sample of FTIR spectroscopy from RAP mastics incorporating various RJs. Relying on their standard deviations, the region from 1600 to 400 cm − 1 reflects the majority of the changes due to variation in RAP concentrations, close to the bitumen fingerprint region [ 18 ]. Only this critical region is shown for brevity, although all recorded wavenumbers will be considered during the classification and pattern recognition tasks. In addition, the spectroscopical absorption of RJs, mineral filler resources, and Asphamin are depicted in Fig. 3 d-f. The spectral data in this figure shows that higher RAP content enhanced absorption intensity within 600 − 400 cm − 1 and 1700 − 1150 cm − 1 ranges while reducing it in 1150 − 800 cm − 1 . Comparing these ranges, the absorption peaks within 1000 − 400 cm − 1 are likely attributed to changes in the mineral composition of bituminous mastics, whereas wavenumbers from 1700 − 1400 cm − 1 reflect the chemical changes in the carbonyl, amides, and aromatic bands. Notably, the sulfoxide band, typically around 1030 cm − 1 , may be obscured by the mineral fillers' interface. Additionally, the presence of RJs in RAP mastics influences absorption intensities, which are not visible in this figure. Figure 3 d shows Sylvaroad with a prominent peak at 1739 cm − 1 , indicating high ester content. Esters are recognized for their ability to interact with oxidized compounds in aged bitumen, reducing their concentration. The 1370 cm − 1 peak in Storflux indicates symmetric bending of methyl (CH₃) groups, highlighting a higher presence of aliphatic components in Storflux compared to Sylvaroad. Aliphatic hydrocarbons in rejuvenators help restore hydrocarbon loss from bitumen aging, balancing polar and non-polar components in the bitumen structure. To identify outliers within the dataset, a modified Z-score test, equipped with Median Absolute Deviation (MAD) as outlined in [ 29 ], was conducted using Eq. 1 . In which, x i represents the data point (absorbance at each wavenumber) and the threshold of |Z modified | > 3.0 is controlled to detect anomalies. With five repetitions performed for each sample, the modified Z-score assessed the variability within the measurements: $$\:{Z}_{modified}=\frac{0.6745\times\:({X}_{i}-Median\:of\:all\:spectra)}{Median\left(\left|{X}_{i}-Median\right|\right)}$$ 1 Figure 4 . Z-scores Box plots of FTIR spectroscopy measurements to identify repetitions outliers: (a) neat bitumen-unrejuvenated, (b) neat bitumen-Sylvaroad, (c) neat bitumen-Storflux, (d) Sasobit modified bitumen-unrejuvenated, (e) Sasobit modified bitumen-Sylvaroad, (f) Sasobit modified bitumen-Storflux, (g) Asphamin modified bitumen-unrejuvenated, (h) Asphamin modified bitumen-Sylvaroad, and (i) Asphamin modified bitumen-Storflux 4-2- Chemical markers reflecting the presence of RJs The dataset was initially analyzed to distinguish between the samples according to RJ types, resulting in a classification involving labels of control mastic, unrejuvenated, Storflux, and Silvaroad. The dimensionality of the dataset was reduced by applying the algorithm detailed in the reference [ 2 , 3 , 24 ]. The transformation of the dataset into a new coordinate system generated 17 LVs, determined by plotting a Scree plot illustrating the cumulative variance explained versus the number of LVs. Processing the LVs—which were generated based on different derivation levels—using LDA classified the dataset into distinct classes, as presented in Fig. 5 . It is worth recalling that the present research adopted a multilevel approach to conduct multivariate discriminant analysis to identify influential chemical components while evaluating strong, prominent, and subtle peaks. For classification purposes, 5-fold cross-validation was employed, yielding a mean accuracy of 0.94, with a precision of 0.97, recall of 0.96, and F1-score of 0.96. In this context, precision reflects the proportion of correctly labeled samples among those predicted by the model, recall indicates the model's ability to identify all relevant instances in each class, and the F1-score, as a harmonic mean of precision and recall, reflects the model's effectiveness in both identifying the population of each class and labeling them correctly. The only failure that the model faced in identifying the labels occurred in distinguishing Storflux and Sylvaroad. Classification based on the first two LDA components (capturing over 90% of system variability) showed similar chemical characteristics between the two rejuvenated mastic classes and the control. However, classification using LDA components 1 and 3 positioned control and rejuvenated mastics near zero for the 3rd LDA component, while distinguishing between the two RJ types. This distinction was controlled by wavenumbers with smaller normalized VIP scores. Throughout this classification, normalized VIP scores were calculated and visualized in Figs. 5 c, 5 f, and 5 i, highlighting influential wavenumbers across derivation levels. Interestingly, the C = O stretching vibration of carbonyl functional groups at ~ 1745 cm − 1 exhibited the highest normalized VIP scores. To identify key spectral regions governing multilevel classification, the highlighted wavenumbers in Figs. 5 c, 5 f, and 5 i were projected onto a sample spectrum in Fig. 6 a and subsequent subfigures detailing their corresponding regions. This figure highlights two key observations: (i) in Figs. 6 c–d, higher RAP contents significantly altered absorption intensities; (ii) Fig. 6 e reveals changes in the Sylvaroad-associated absorption peak, where higher RAP content—and subsequently higher RJ content—intensified absorption, which can be inferred as residue RJs due to partial blending of RAP binder and RJ. Using OPUS software, integrals were calculated for the spectral ranges 590 − 555 cm − 1 , 1330 − 1275 cm − 1 , 1370 − 1360 cm − 1 , and 1770 − 1725 cm − 1 . Beyond their chemical significance and the trends shown in Fig. 7 , two key aspects were considered for feature selection: (i) the contribution of these bands to the classification in Fig. 5 and (ii) their sensitivity to variables such as RAP content, RJs, and warm mix additives. To assess feature importance, four ML algorithms—random forest, logistic regression, XGBoost, and CatBoost—were applied to classify the dataset based on the integrals. The results consistently ranked the absorption bands in order of importance:1758 − 1715 cm − 1 > 1370 − 1360 cm − 1 > 1330 − 1275 cm − 1 > 590 − 555 cm − 1 . The dependency of these spectral ranges on the testing variables was further analyzed using the same ML algorithms. The results indicated that the integral values of 1758 − 1715 cm − 1 and 1370 − 1360 cm − 1 were primarily influenced by RJs rather than RAP content or warm mix additives. In contrast, the integral values of 1330 − 1275 cm − 1 were more dependent on RAP content. The analysis also suggested the reflection of warm mix additives within these ranges. From a chemical perspective, the 1758 − 1715 cm − 1 range corresponds to carbonyl stretching in carboxylic acids (-COOH) [ 31 ] and C = O stretching in esters [ 32 ], both of which form during aging. In rejuvenated bituminous mixtures, esters in this range may indicate resistance to aging and the effectiveness of rejuvenation [ 33 ]. The spectral results of Sylvaroad, shown in Fig. 3 d, exhibit a significant peak within this range, suggesting that the prominent peaks in Fig. 6 d correspond to the presence of this specific RJ. The peaks at 1370 − 1360 cm − 1 correspond to C-H bending vibrations in hydrocarbons, specifically in methyl (CH 3 ) and methylene (CH 2 ) groups. Methyl groups are primarily found in saturated and aromatic hydrocarbons, while methylene groups are more common in aliphatic hydrocarbons. This range may indicate the influence of recycling agents, where increased intensity suggests a potential restoration of aliphatic components in bitumen, enhancing flexibility—consistent with the feature importance findings. A comparison of the integrals in Fig. 7 b highlights the distinction between Storflux-modified samples and others. Additionally, the absorption intensity of various RJs in Fig. 3 d exhibits a characteristic peak at 1370 cm − 1 , confirming differences in their integral values. The 1330 − 1275 cm − 1 range corresponds to C-O stretching vibrations in hydroxyl (-OH) groups, commonly found in alcohols or phenolic compounds [ 34 ]. The oxidation of bitumen in RAP binders likely increases hydroxyl group concentrations as RAP content rises. This aligns with the feature importance results, where RAP content was identified as a key factor influencing this range (Fig. 6 b). However, Fig. 7 a indicates minimal changes in integral values after RJ incorporation, suggesting these oxidative changes remain largely irreversible, consistent with findings in the literature. Since this range does not significantly differentiate mastics before and after rejuvenation, it primarily distinguishes control mastics from RAP mastics. The absorption at 590 − 555 cm − 1 in the zero-order dataset reflects metal-oxygen bending, attributed to phosphate (PO 4 3− ) compounds, as indicated by Table 7.5 in [ 30 ]. The increasing intensity in this range with higher RAP content suggests a notable contribution from phosphate compounds, likely originating from RAP fillers. A comparison of FTIR spectra for RAP and basalt fillers (Fig. 3 e) further supports this, with a pronounced absorption near 500 cm − 1 reinforcing the presence of phosphate-related metal oxides introduced by RAP fillers into the mastic. 4-3- Chemical markers reflecting the effectiveness of Sylvaroad Upon distinguishing between various RJs in RAP mastics, their mechanisms in mitigating chemical changes induced by RAP binders remain unclear. This section identifies chemical markers reflecting Sylvaroad’s effectiveness in mitigating RAP binder-induced changes. Key markers distinguishing RAP mastics before and after rejuvenation are examined, where those losing significance post-rejuvenation indicate Sylvaroad’s influence. Figures 8 a, 8 e, and 8 i show classification before rejuvenation across different derivation orders, while Figs. 8 c, 8 g and 8 k depict Sylvaroad-rejuvenated mastics. Normalized VIP scores from zero-order datasets suggest that Syvlaroad’s effect is not strongly reflected in major absorption peaks, except at 2786 cm − 1 , linked to C–H stretching in aliphatic hydrocarbons, possibly indicating a subtle oxidation effect. In the first derivative dataset, peaks at 567 cm − 1 , 803 cm − 1 , and 1160 cm − 1 appear before and after rejuvenation, but Sylvaroad notably reduces the significance of 2748 cm − 1 . The second derivative dataset does not clearly distinguish Sylvaroad’s impact based on normalized VIP scores. To further examine these wavenumbers, they are projected onto a spectral sample representing the effect of increasing RAP concentrations. This projection, shown in Fig. 9 , highlights absorption regions capturing the influence of RAP variation and, presumably, Sylvaroad. Table 2 summarizes these ranges, detailing their derivation level, corresponding functional groups, and interpretation. Table 2 suggests that the presence of RAP fractions in both control and Sylvaroad-rejuvenated mastics can be assessed by evaluating the integrals of specific absorption ranges. The 489 − 446 cm − 1 range (~ Figure 9 b), associated with the Si-O band, can be a marker of minerals such as fillers or organic compounds in Asphamin. The 590 − 555 cm − 1 range (~ Figure 9 c) corresponds to phosphate (PO 4 3− ) compounds, indicating potential mineral compounds. The 650 − 610 cm − 1 range (~ Figure 9 d), linked to aromatic out-of-plane vibrations of hydrocarbons or sulfates, and the 810 − 785 cm − 1 range (~ Figure 9 e), associated with out-of-plane C-H bending vibrations, particularly in meta-distributed benzene rings per [ 30 ], are also key indicators. Typically, aging reduces absorption intensity in the 810 − 785 cm − 1 range [ 35 ], Fig. 9 e, as seen in the comparison between RAP and virgin binders in Fig. 3 d. However, the observed peak may also originate from recovered RAP fillers, with peaks at 770 cm − 1 , 789 cm − 1 , and 879 cm − 1 in Fig. 3 e. Specifically, the 789 cm − 1 peak indicates the quartzite nature of RAP fillers, while the 879 cm − 1 peak points to the carbonate [ 36 ]. These findings suggest that increasing RAP content creates a local peak in this range, useful for distinguishing RAP contents based on mineral components. The 1300 − 1160 cm − 1 range (~ Figure 9 f) likely reflects various chemical compounds, including S = O stretching in sulfonates, sulfates, and sulfoxides, as well as C-O stretching in ethers and carboxylic acid derivatives [ 30 , 34 ]. Mirwald et al. [ 37 ] and Primerano et al. [ 38 ] attributed the increased absorption in this range to the overall increase in bitumen’s polarity due to oxidative aging. This increase in absorption, correlates with higher RAP content, indicating the migration of oxidized components from the RAP binder to the blended binder, consistent with findings in [ 21 ]. The upward trend in absorption suggests a higher concentration of oxidized compounds in RAP mastics. Finally, the 1770 − 1725 cm − 1 range (~ Figure 9 g) represents C = O stretching, a known marker of oxidative aging, even may also be attributed to Sylvaroad, which shows a strong absorption at 1739 cm − 1 , as depicted in Fig. 3 d. Table 2 Significant wavenumbers based on classification are shown in Fig. 8 Proposed range for integration Zero-order spectrum 1st derivative 2nd derivative Functional group Interpretation 489 − 446 cm − 1 ✔ ✘ ✘ Si–O bending or Al–O vibrations Mineral filler contribution (e.g., aggregates, additives like Asphamin) 590 − 555 cm − 1 ✔ ✘ ✘ Metal-oxygen bending Phosphate (PO 4 3− ) compounds from RAP fillers 650 − 610 cm − 1 ✘ ✔ ✔ Aromatic C–H out-of-plane bending Aromatic ring structures from aged bitumen in RAP 810 − 785 cm − 1 ✘ ✔ ✔ Aromatic C–H out-of-plane bending Aged bitumen’s aromatic character, possibly modified by rejuvenators 1300 − 1160 cm − 1 ✘ ✔ ✔ C–O stretching Formation of oxygenated compounds such as alcohols and esters 1720 − 1670 cm − 1 ✘ ✘ ✔ C = O stretching (carbonyls, ketones, esters) Oxidation products presumably from RAP binder aging 1770 − 1725 cm − 1 ✘ ✘ ✔ Carbonyl groups such as esters, aldehydes, and ketones The marker of Sylvaroad 4-4- Chemical markers reflecting the effectiveness of Storflux Similarly, the subset of RAP mastics incorporating Storflux was analyzed using the same approach, leading to a multilevel classification based on different derivation levels (Fig. 10 ). Notably, this classification revealed that no significant pattern emerged from the zero-order dataset, where the distinction shown in Fig. 10 a primarily captured noise, resulting in inconclusive findings. Evaluation of the first derivative results, Fig. 10 b, identified wavenumbers at 656 cm − 1 , 1729 cm − 1 , and 3249 cm − 1 as significant. The second derivative analysis, Fig. 10 c, highlighted wavenumbers at 658 cm − 1 and 3077 cm − 1 as important. These frequencies suggest that the distinction between mastic classes is primarily governed by carbonyl compounds, aromatic ring structures, and O–H stretching vibrations. The normalized VIP scores corresponding to each classification are presented alongside the classification results in Figs. 10 b, 10 d, and 10 f. Projecting these wavenumbers onto a spectral sample (Fig. 11 ) revealed two key absorption regions likely contributing to the observed distinctions: 650 − 610 cm − 1 and 1670–1720 cm − 1 . The former region was identified from both the first and second derivatives, while the latter was distinguished through first-derivative analysis. The 650–610 cm − 1 range can be attributed to aromatic out-of-plane bending vibrations in hydrocarbons and S-O bending vibrations in inorganic sulfates, as indicated by Tables 2.6, 6.5, and 7.1 in [ 30 ]. The reduction in intensity absorption within this range, due to increased RAP content, aligns with the spectral results of RAP fillers shown in Fig. 3 e. The absorption within the 1670–1720 cm − 1 range captures the accumulation of carbonyl compounds due to increased RAP concentrations. The prominence of these two absorption regions, independent of the wavenumbers identified in unrejuvenated mastics, suggests that they serve as chemical markers correlated with oxidation and RAP mineral fillers, reflecting higher RAP contents. Furthermore, these findings indicate that Storflux played a role in restoring the significant maltene loss associated with aging. Although higher RAP contents are typically identified by an increase in sulfoxide and/or carbonyl compounds, the interaction between basalt filler and the accumulation of Si-containing compounds likely masked oxidation markers within the sulfoxide region for the mastics tested in this study. Instead, evaluating the carbonyl functional groups, as shown in Fig. 11 c, indicates that higher RAP contents lead to the accumulation of oxidative compounds. 4-5- Pattern recognition Multivariate discriminant analysis identified several absorption ranges that distinguish mastic classes with varying RAP concentrations before and after rejuvenation. The analysis revealed the formation of distinctive differentiation patterns influenced by Sylvaroad and Storflux. Notably, the incorporation of Storflux, unlike Sylvaroad, significantly altered the differentiation pattern, as reflected in the normalized VIP scores computed at various derivation levels. This section aims to refine the observed patterns and elucidate the nonlinear relationships shaping these classifications. To achieve this, spectral datasets underwent integration within the identified ranges using OPUS software. Among these, the 489 − 446 cm − 1 and 590 − 555 cm − 1 regions could potentially serve as mineral composition markers within the blended mastics. Since each mixture contained three distinct mineral sources—virgin filler, RAP mineral particles, and Asphamin—the effect of increasing RAP content was evaluated separately through the corresponding integral values, as depicted in Fig. 12 . Two key criteria were considered in this evaluation: (1) changes in integral values due to increasing RAP concentrations and (2) the distribution of recorded data points within each box, representing the effects of three RJ inclusion levels (unrejuvenated, Sylvaroad, and Storflux). The data distribution analysis indicates that RJ inclusion had a negligible effect on the integral values of the 590–555 cm − 1 region and a relatively minor influence on the 489 − 446 cm − 1 range. Consequently, these two regions appear to primarily capture the inclusion of higher RAP contents, serving as chemical markers of mineral composition. Further assessment of RAP content variations revealed a progressive increase in integral values for mastics containing neat bitumen and Sasobit, whereas an inverse trend was observed in mastics with Asphamin. Given that Asphamin exhibits a pronounced peak at 465 cm − 1 , it contributes to a notable increase in absorption intensity and integral values. However, as RAP content increases, the intensity difference between Asphamin mastics and other variants diminishes. This suggests that Asphamin introduces Al–O absorption, likely from zeolite, while RAP contributes to Si–O absorption. The interaction between these two mineral phases at higher RAP concentrations likely leads to spectral flattening, reducing overall integral values. This phenomenon may indicate a redistribution effect, wherein Al–O and Si–O vibrations disperse, leading to a more uniform spectral response–an observation that requires further validation. Regarding the 590 − 555 cm − 1 integral values, an increase in RAP content resulted in a gradual decline, regardless of the type of warm mix additives used. The lower absorption intensities in this range, compared to those of basalt fillers, suggest a masking effect on sulfate-related absorptions. To evaluate changes in the composition of aromatic C–H out-of-plane bending, observed in the 650 − 610 cm − 1 and 810 − 785 cm − 1 regions, integral values were plotted in Fig. 13 . The results indicate a continuous decline in absorbance intensities within the 650 − 610 cm − 1 range with increasing RAP concentrations. A comparison of subfigures corresponding to different rejuvenation levels suggests a slight but noticeable influence of rejuvenation; however, the overall trend remained unchanged. Conversely, an increase in RAP concentration led to a gradual rise in integral values within the 810 − 785 cm − 1 range, though at a slower rate. These trends are chemically consistent with the interpretation of the associated spectral regions, as discussed earlier. On the other hand, this figure highlights two key observations: (i) the variability introduced by warm mix additives, which is reflected in the scatter of recorded data points within each box and indicates their influence within each mastic class; and (ii) the negligible contribution of RJs in mitigating aging-related compounds that migrated from the RAP binder into the blended binder. The latter point suggests that the spectral features identified correspond to aromatic C–H out-of-plane bending primarily capturing oxidation-induced chemical changes that remain irreversible despite rejuvenation efforts. To assess the regions indicative of oxygen-containing markers, the integral values corresponding to the 1300 − 1160 cm − 1 and 1720 − 1670 cm − 1 regions were plotted in Fig. 14 . The 1300 − 1160 cm − 1 range, associated with the formation of oxygenated compounds such as esters and alcohols, exhibited a continuous increase in integral values, irrespective of the rejuvenation technique employed. However, a detailed comparison of data at the same RAP concentration level revealed that Sylvaroad led to a slight reduction in integral values, particularly at higher RAP concentrations. Similarly, the integral values within 1720 − 1670 cm − 1 , which reflect the formation of carbonyl groups, showed a gradual increase, although at a slower rate. This trend suggests the migration of oxygenated compounds from the RAP binder into the blended binder, with Sylvaroad exerting a rather moderate mitigating effect on this transformation. The absorption region at 1770 − 1725 cm − 1 has previously been attributed to the characteristic absorption peak of Sylvaroad (Fig. 3 d). The computed integral values, as presented in Fig. 15 , indicate a distinct separation of mastics containing Sylvaroad from the others, regardless of the type of warm mix additives used. 4-6- Irreversible components post-rejuvenation Previous sections examined chemical components associated with RJs and those reflecting varying RAP concentrations after rejuvenation. However, a key question remains: which aging-induced chemical changes persist in post rejuvenation? To address this, the feature extraction results, based on PLSR-LDA approach, were employed to identify chemical markers that distinguish mastic classes both before and after rejuvenation. This approach focuses on detecting features that were significant in unrejuvenated mastics and remain relevant even after rejuvenation, highlighting irreversible chemical changes. Initially, normalized VIP scores, based on the first derivative dataset, were obtained from analyses of mastics before rejuvenation and mapped against those post-rejuvenation—in which the post-rejuvenated class involved both RJs. Figure 16 a presents the projection of normalized VIP scores from post-rejuvenation classification onto the pre-rejuvenation dataset. Figure 16 b provides an enlarged view of wavenumbers with higher normalized VIP scores in both pre- and post-rejuvenation analyses. This indicates that the selected wavenumber range retains valuable information for distinguishing between mastic classes in both conditions. In other words, these ranges play a critical role in differentiating mastics with varying RAP percentages, regardless of rejuvenation. Therefore, these wavenumber ranges may represent irreversible chemical markers that persist after rejuvenation. The highlighted wavenumbers and their associated absorption ranges in Fig. 16 b closely align with those identified in previous sections. As previously discussed, the wavenumbers at 642 cm − 1 and 768 cm − 1 are linked to mineralogical changes in RAP mastics, while 1160 cm − 1 is associated with oxidation phenomena. For unrejuvenated RAP mastics, similar absorption patterns were observed as in Fig. 8 , though with varying intensity levels, which are omitted here for brevity. To assess the correlation and consistency of absorption intensities across these ranges, integral values before and after rejuvenation were compared (Fig. 17 ). This figure presents the average integral values for the respective ranges alongside their standard deviations, evaluating the impact of RJs on these absorption regions. Two key observations can be drawn from this figure: (i) variations in standard deviations and the (ii) minimal impact of rejuvenation on absorption intensity. The data illustrated in Fig. 17 further validates the reliability of the PLSR-LDA model. As anticipated, the impact of rejuvenation on absorption regions related to the mineralogical characteristics of mastics was minimal. However, Fig. 17 a reveals relatively high standard deviations for integral values, likely due to the influence of warm mix additives, particularly Asphamin, which may alter the mineral composition of mastics by introducing metals and mineral oxides. In contrast, Fig. 17 c indicates that the absorption range of 1300–1160 cm − 1 shows negligible variation with rejuvenation, accompanied by small standard deviation values. This range can therefore be identified as indicative of irreversible changes post-rejuvenation. The magnitude of the integrals for these wavenumber ranges also provides a basis for distinguishing mastics based on RAP content, both before and after rejuvenation. However, other regions and functional groups may also influence the blending of RAP binder with virgin bitumen. These effects potentially masked by the more prominent influence of mineral components. To overcome this limitation, further evaluations at the binder scale—by isolating the blended binder from recycled mixtures and excluding mineral fillers—could offer valuable insights. These investigations may provide a more refined understanding of irreversible chemical markers post-rejuvenation. 5 Conclusions and recommendations This study explored the rejuvenation mechanisms in bituminous RAP mastics using FTIR spectroscopy and multivariate discriminant analysis. The goal was to examine the chemical between RAP binder, virgin bitumen, warm mix additives, and RJs to uncover the the underlying mechanisms of RAP binder rejuvenation. To avoid analytical biases from inconsistencies in the filler-to-binder ratio, mastics with varying RAP concentrations were produced using a precise mix design, maintaining a constant volumetric value. BTSV testing was conducted to identify the optimal RJ content, followed by a hybrid PLSR-LDA analysis of FTIR results, leading to the following findings: The multilevel approach employed in this study, combining PLSR-LDA with normalized VIP scores, successfully identified key FTIR frequencies responsible for differentiating RAP mastics across various derivative levels. Classification and feature extraction tasks to explore the presence of RJs in RAP mastics revealed the dominant role of absorption bands at 1330-1275 cm-1 (~C-O stretching vibrations in hydroxyl groups), 1370-1360 cm-1 (~C-H bending vibrations in hydrocarbons), and 1758-1715 cm-1 (~carbonyl stretching in carboxylic acids and esters). Among these ranges, the first is attributed to oxidative chemical changes, distinguishing control mastics, unrejuvenated RAP mastics, and rejuvenated mastics, reflecting irreversible chemical changes. The latter two ranges correspond to specific chemical markers associated with the RJs used. Investigation of the RJs impacts, three categories of chemical markers were identified as reflecting the changes post-rejuvenation: ranges linked to mineralogical characteristics (i.e., 489-446 cm -1 and 590-555 cm -1 ), ranges attributed to aromatic markers (i.e., 650-610 cm -1 and 810-785 cm -1 ), and ranges reflecting the changes within the oxygen containing frequencies (i.e., 1300-1160 cm -1 and 1720-1670 cm -1 ). This observation suggests that varying RAP contents in mastics post-rejuvenation were identified based on both fillers’ mineralogical characteristics and aging-induced changes. To identify aging-induced chemical markers that remain irreversible after rejuvenation, three absorption ranges were identified: 650-610 cm -1 , 785-770 cm -1 , and 1300-1160 cm -1 . Evaluating their integral values before and after RJ incorporation showed consistent values, indicating the critical role of these regions in distinguishing mastic classes and their stability post-rejuvenation. Among these, the 1300-1160 cm -1 range is linked to oxidation, reflecting irreversible chemical changes post-rejuvenation. Declarations Acknowledgment The authors wish to extend their appreciation to the Alexander von Humboldt Foundation for their support of this project through the Georg Forster Postdoctoral Fellowship awarded to the first author. Data availability Data will be made available on request. References Mocelin DM, Isied MM, da Costa RF, Castorena C (2024) Availability adjusted mix design method as a tool to mitigate the adverse effects of RAP on the performance of asphalt mixtures. Constr Build Mater 422. https://doi.org/10.1016/j.conbuildmat.2024.135813 Motevalizadeh SM, Mollenhauer K, Wetekam J (2024) FTIR spectroscopy and multivariate discriminant analysis for classifying bituminous mastics: Exploring aging states and mastic composition. Constr Build Mater 438. https://doi.org/10.1016/j.conbuildmat.2024.137188 Motevalizadeh SM, Mollenhauer K (2024) Exploration of chemical changes in bituminous mastics induced by aging: insights from FTIR spectroscopy, DSR measurements, and machine learning. Int J Pavement Eng. https://doi.org/https://doi.org/10.1080/10298436.2024.2418927 Tauste R, Moreno-Navarro F, Sol-Sánchez M, Rubio-Gámez MC (2018) Understanding the bitumen ageing phenomenon: A review. Constr Build Mater 192:593–609. https://doi.org/10.1016/j.conbuildmat.2018.10.169 Behnood A, Modiri Gharehveran M (2019) Morphology, rheology, and physical properties of polymer-modified asphalt binders. Eur Polym J 112:766–791. https://doi.org/10.1016/j.eurpolymj.2018.10.049 Behnood A (2019) Application of rejuvenators to improve the rheological and mechanical properties of asphalt binders and mixtures: A review. J Clean Prod 231:171–182. https://doi.org/10.1016/j.jclepro.2019.05.209 Nam T, Taylor A, Willis R (2012) Effect of rejuvenator on performance properties of hma mixtures with high rap and ras contents Martin Zaumanis,Rajib RF (2013) B.Mallick, Evaluation of Rejuvenator’s Effectiveness with Conventional Mix Testing for 100% RAP Mixtures. Transp Res Board 1–14 Nahar SN, Qiu J, Schmets AJM, Schlangen E, Shirazi M, Van De Ven MFC, Schitter G, Scarpas A (2014) Turning back time: Rheological and microstructural assessment of rejuvenated bitumen. Transp Res Rec 2444:52–62. https://doi.org/10.3141/2444-06 Ali AW, Mehta YA, Nolan A, Purdy C, Bennert T (2016) Investigation of the impacts of aging and RAP percentages on effectiveness of asphalt binder rejuvenators. Constr Build Mater 110:211–217. https://doi.org/10.1016/j.conbuildmat.2016.02.013 Abdelaziz A, Epps Martin A, Masad E, Arámbula Mercado E, Kaseer F (2022) Effects of ageing and recycling agents on the multiscale properties of binders with high RAP contents. Int J Pavement Eng 23:1248–1270. https://doi.org/10.1080/10298436.2020.1797736 Holleran I, Masad E, Holleran G, Wubulikasimu Y, Malmstrom J, Wilson DJ (2021) Nanomechanical mapping of rejuvenated asphalt binders. Road Mater Pavement Des 22:2478–2497. https://doi.org/10.1080/14680629.2020.1771406 Cavalli MC, Partl MN, Poulikakos LD (2019) Effect of ageing on the microstructure of reclaimed asphalt binder with bio-based rejuvenators, Road Mater. Pavement Des 20:1683–1694. https://doi.org/10.1080/14680629.2019.1594049 Menapace I, Garcia Cucalon L, Kaseer F, Arámbula-Mercado E, Epps Martin A, Masad E, King G (2018) Effect of recycling agents in recycled asphalt binders observed with microstructural and rheological tests. Constr Build Mater 158:61–74. https://doi.org/10.1016/j.conbuildmat.2017.10.017 Yu X, Zaumanis M, Dos Santos S, Poulikakos LD (2014) Rheological, microscopic, and chemical characterization of the rejuvenating effect on asphalt binders. Fuel 135:162–171. https://doi.org/10.1016/j.fuel.2014.06.038 Abdelaziz A, Masad E, Epps Martin A, Mercado EA, Bajaj A (2021) Multiscale Characterization of Aging and Rejuvenation in Asphalt Binder Blends with High RAP Contents. J Mater Civ Eng 33. https://doi.org/10.1061/(asce)mt.1943-5533.0003910 Pipintakos G, Soenen H, Ching HYV, Vande Velde C, Van Doorslaer S, Lemière F, Varveri A (2021) Van den bergh, Exploring the oxidative mechanisms of bitumen after laboratory short- and long-term ageing. Constr Build Mater 289. https://doi.org/10.1016/j.conbuildmat.2021.123182 Mirwald J, Werkovits S, Camargo I, Maschauer D, Hofko B, Grothe H (2020) Understanding bitumen ageing by investigation of its polarity fractions. Constr Build Mater 250. https://doi.org/10.1016/j.conbuildmat.2020.118809 Mastoras F, Varveri A, van Tooren M, Erkens S (2021) Effect of mineral fillers on ageing of bituminous mastics. Constr Build Mater 276. https://doi.org/10.1016/j.conbuildmat.2020.122215 Bajaj A, Epps Martin A, King G, Glover C, Kaseer F (2020) Arámbula-Mercado, Evaluation and classification of recycling agents for asphalt binders. Constr Build Mater 260. https://doi.org/10.1016/j.conbuildmat.2020.119864 Ma L, Varveri A, Jing R, Erkens S (2023) Chemical characterisation of bitumen type and ageing state based on FTIR spectroscopy and discriminant analysis integrated with variable selection methods. Road Mater Pavement Des. https://doi.org/10.1080/14680629.2023.2181008 Weigel S, Stephan D (2018) Differentiation of bitumen according to the refinery and ageing state based on FTIR spectroscopy and multivariate analysis methods. Mater Struct Constr 51. https://doi.org/10.1617/s11527-018-1252-6 Weigel S, Wetekam J, Mollenhauer K (2023) Identification and classification of PAH in asphalt binders with FTIR spectroscopy and multivariate analysis methods. Fuel 337. https://doi.org/10.1016/j.fuel.2022.126845 Motevalizadeh SM, Mollenhauer K (2024) Use of multivariate clustering analysis to investigate the physicochemical interactions in bitumen mastics using micromechanical modeling and FTIR spectroscopy. Constr Build Mater 448:138230. https://doi.org/10.1016/j.conbuildmat.2024.138230 Gundla A, Underwood S (2017) Evaluation of in situ RAP binder interaction in asphalt mastics using micromechanical models. Int J Pavement Eng 18:798–810. https://doi.org/10.1080/10298436.2015.1066003 Walther A, Büchler S, Cannone Falchetto A, Wang D, Riccardi C, Wistuba MP (2019) Experimental investigation on asphalt mixtures prepared with reclaimed asphalt pavement and rejuvenators based on the BTSV method. Road Mater Pavement Des 20:1695–1708. https://doi.org/10.1080/14680629.2019.1594053 Alisov A, Riccardi C, Schrader J, Cannone Falchetto A, Wistuba MP (2020) A novel method to characterise asphalt binder at high temperature. Road Mater Pavement Des 21:143–155. https://doi.org/10.1080/14680629.2018.1483258 Schrader J, Wistuba MP, Falchetto AC, Riccardi C, Alisov A (2020) A new binder-fast-characterization-test using a dynamic shear rheometer and its application for rejuvenating reclaimed asphalt binder. J Test Eval 48. https://doi.org/10.1520/JTE20180893 Yaro AS, Maly F, Prazak P, Maly K (2024) Outlier Detection Performance of a Modified Z-Score Method in Time-Series RSS Observation With Hybrid Scale Estimators. IEEE Access 12:12785–12796. https://doi.org/10.1109/ACCESS.2024.3356731 Smith B (1998) Infrared spectral interpretation: a systematic approach. https://doi.org/https://doi.org/10.1201/9780203750841 Bakhtiari MT (2015) Role of Sodium Hydroxide in Bitumen Extraction: Production of Natural Surfactants and Slime Coating, 206 Xiaolong Zou XW, Xie H, Li J, Jing H, Li H, Li Z (2024) Innovative application of coffee grounds oil as an asphalt modifier: Extraction, preparation, and rheological properties, Case Stud. Constr Mater 21. https://doi.org/https://doi.org/10.1016/j.cscm.2024.e04012 Loise V, Caputo P, Porto M, Calandra P, Angelico R, Rossi CO (2019) A review on Bitumen Rejuvenation: Mechanisms, materials, methods and perspectives. Appl Sci 9. https://doi.org/10.3390/app9204316 Hesse M, Meier H, Zeeh B (2005) –7 Spektroskopische Methoden in der organischen Chemie Auflage), 7th ed., Georg Thieme, Stuttgart, 2005 Ren S, Liu X, Lin P, Jing R, Erkens S (2022) Toward the long-term aging influence and novel reaction kinetics models of bitumen. Int J Pavement Eng 24. https://doi.org/10.1080/10298436.2021.2024188 Liu X, Colman SM, Brown ET, Minor EC, Li H (2013) Estimation of carbonate, total organic carbon, and biogenic silica content by FTIR and XRF techniques in lacustrine sediments. J Paleolimnol 50:387–398. https://doi.org/10.1007/s10933-013-9733-7 Mirwald J, Nura D, Hofko B (2022) Recommendations for handling bitumen prior to FTIR spectroscopy. Mater Struct Constr 55. https://doi.org/10.1617/s11527-022-01884-1 Hofer K, Mirwald J, Maschauer D, Grothe H, Hofko B (2022) Influence of selected reactive oxygen species on the long-term aging of bitumen. Mater Struct Constr 55. https://doi.org/10.1617/s11527-022-01981-1 Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6342505","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":436170076,"identity":"089b23f7-5f1e-4894-bbaf-8deb24895874","order_by":0,"name":"Mohsen Motevalizadeh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABK0lEQVRIie3RsUrDQBjA8TsOLoNNu0YC9gmEFKEQqORVehSaqXuGUA4C6VJ1TaAPkSI4XzhIlzzAgR0iBacOmYqiBC9WcTBR3BzuD8nwcb8cdwFApfrPdZF8dQA7AwBSSOuRBoDRtpzJB3+Qiy+CfiPgSEi9+kdyTk/IrvR4H2vaQ7n3tm5yH1AY+XOnhwAU5XcyZPqtxXI+CBGYnK7yx1myTSlMMk7iACA7aiR3RhpOoSTM1EM+SwRZPBeYjS0OsNlpI9XUCREMXvSKu5YgFBbV3HFq8tpG6IiECGFTp3z8TtbyCwmSpOH4Qy7PkmejSYgwtlcZH8Q1ia85iTgM7GUD2VytC883Lm96m53Y+7zfFW4Blwd5Y4sgFU8N26CG2WfH36NSqVSqv/cGjSdt1UNjqOoAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-8304-6271","institution":"University of Kassel","correspondingAuthor":true,"prefix":"","firstName":"Mohsen","middleName":"","lastName":"Motevalizadeh","suffix":""},{"id":436170077,"identity":"53a103cf-d47f-4ea3-ae99-c0e64342209f","order_by":1,"name":"Konrad Mollenhauer","email":"","orcid":"","institution":"University of Kassel","correspondingAuthor":false,"prefix":"","firstName":"Konrad","middleName":"","lastName":"Mollenhauer","suffix":""}],"badges":[],"createdAt":"2025-03-31 07:26:18","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6342505/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6342505/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":79661368,"identity":"86fcfa2a-3006-4c8e-b045-d53c5a8d929c","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":174482,"visible":true,"origin":"","legend":"\u003cp\u003eBTSV testing results on bitumen samples: (a) virgin 50/70 bitumen, (b) RAP recovered binder, (c) RAP binder modified with 10% w/w Sylvaroad, and (d) RAP binder modified with 15% w/w Storflux\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/ea9c8208db28cf4e4a880d92.png"},{"id":79661403,"identity":"c26ba84f-93fe-4723-b5b1-bf9ca0d5eab7","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":80212,"visible":true,"origin":"","legend":"\u003cp\u003eProjecting BTSV testing results to identify the optimum RJ dosages\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/5f840ee75069a9a25afcc380.png"},{"id":79661365,"identity":"18a7ec81-62b8-4111-9937-be503713ef5c","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":621941,"visible":true,"origin":"","legend":"\u003cp\u003eFTIR spectra results of: (a) unrejuvenated mastics (w/o warm mix additives), (b) Storflux mastics (w/o warm mix additives), (c) Sylvaroad mastics (w/o warm mix additives), (d) RJs, (e) fillers, and (f) Asphamin\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/47fc67af8a74ffe6929597a7.png"},{"id":79662007,"identity":"55da5006-d11b-4e44-8e0b-031cfd5d88c6","added_by":"auto","created_at":"2025-04-01 09:49:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":357163,"visible":true,"origin":"","legend":"\u003cp\u003eZ-scores Box plots of FTIR spectroscopy measurements to identify repetitions outliers: (a) neat bitumen-unrejuvenated, (b) neat bitumen-Sylvaroad, (c) neat bitumen-Storflux, (d) Sasobit modified bitumen-unrejuvenated, (e) Sasobit modified bitumen-Sylvaroad, (f) Sasobit modified bitumen-Storflux, (g) Asphamin modified bitumen-unrejuvenated, (h) Asphamin modified bitumen-Sylvaroad, and (i) Asphamin modified bitumen-Storflux\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/72c808977b57d964ebd38884.png"},{"id":79661407,"identity":"edeab479-2b18-409b-a348-9b4017f5b250","added_by":"auto","created_at":"2025-04-01 09:41:16","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":365711,"visible":true,"origin":"","legend":"\u003cp\u003eClassification of RAP mastics based on the RJs used. Subfigures (a-b) show the classification results for the zero-order dataset, (d-e) for the first derivative, and (g-h) for the second derivative. Normalized VIP scores corresponding to each dataset are presented in (c) for the zero-order, (f) for the first derivative, and (i) for the second derivative.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/eeaf65dc4feae236bb4dd754.png"},{"id":79662008,"identity":"72d2a2af-2a04-42ef-82de-b4d75666454c","added_by":"auto","created_at":"2025-04-01 09:49:15","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":902888,"visible":true,"origin":"","legend":"\u003cp\u003eProjection of highlighted wavenumbers in Figure 5: (a) on a sample FTIR spectra, (b) 590-555 cm\u003csup\u003e-1\u003c/sup\u003e, (c) 1330-1275 cm\u003csup\u003e-1\u003c/sup\u003e, (d) 1370-1360 cm\u003csup\u003e-1\u003c/sup\u003e, and (e) 1758-1715 cm\u003csup\u003e-1\u003c/sup\u003e\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/495f57b32cc1503bc6b313b0.png"},{"id":79661384,"identity":"868a3f0c-e552-4095-b96c-34cfdf8c5d71","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":199185,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of wavenumber intensities for mastics with various RJs: (a) 590-555 cm\u003csup\u003e-1\u003c/sup\u003e, (b) 1330-1275 cm\u003csup\u003e-1\u003c/sup\u003e, (c) 1370-1360 cm\u003csup\u003e-1\u003c/sup\u003e, and (d) 1758-1715 cm\u003csup\u003e-1\u003c/sup\u003e\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/28706e27b23fed4d3dac1704.png"},{"id":79662013,"identity":"aed4d350-6c90-4ad1-bf61-34f1e9b16f77","added_by":"auto","created_at":"2025-04-01 09:49:15","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":462325,"visible":true,"origin":"","legend":"\u003cp\u003eClassification results and corresponding normalized VIP scores across different derivation levels for unrejuvenated mastics: (a, b) zero-order, (e, f) first, and (i, j) second derivative levels; for Sylvaroad-modified mastics: (c, d) zero-order, (g, h) first, and (k, l) second derivative levels\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/554cba409d389ad97badfab0.png"},{"id":79661389,"identity":"a75a3fcd-5796-4f62-94f6-d8cebaa7fb28","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":211942,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Projection of key wavenumbers associated with Sylvaroad modification onto the spectroscopic results, with detailed views of the absorption ranges at (b) 489-446 cm\u003csup\u003e-1\u003c/sup\u003e, (c) 590-555 cm\u003csup\u003e-1\u003c/sup\u003e, (d) 650-610 cm\u003csup\u003e-1\u003c/sup\u003e, (e) 810-785 cm\u003csup\u003e-1\u003c/sup\u003e, (f) 1300-1160 cm\u003csup\u003e-1\u003c/sup\u003e, and (g) 1770-1725 cm\u003csup\u003e-1\u003c/sup\u003e\u003c/p\u003e","description":"","filename":"image9.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/b1030949c0e3638cf9a32600.png"},{"id":79661393,"identity":"0ec48e80-dd66-44f0-94a2-7e78092c1e0e","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":210781,"visible":true,"origin":"","legend":"\u003cp\u003eClassification results and corresponding normalized VIP scores across different derivation levels for Storflux-modified mastics: (a-b) zero-order, (c-d) first, and (e-f) second derivative levels\u003c/p\u003e","description":"","filename":"image10.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/4fa3e8facc0d001bec23b276.png"},{"id":79661369,"identity":"55d7f4ce-dded-4a57-a349-214dbe9cd04c","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":244072,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Projection of key wavenumbers associated with Storflux modification onto the spectroscopic results, with detailed views of the absorption ranges at (b) 650-610 cm\u003csup\u003e-1\u003c/sup\u003e, and (c) 1720-1670 cm\u003csup\u003e-1\u003c/sup\u003e\u003c/p\u003e","description":"","filename":"image11.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/a06c408ac4620059ff7d9df1.png"},{"id":79661372,"identity":"d5bb4817-7f35-41f3-abbb-46335b7addba","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":332714,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of integral values for two spectral ranges across different mastic compositions: (a–c) 489–446 cm\u003csup\u003e-1\u003c/sup\u003e range for mastics containing (a) neat bitumen, (b) Sasobit, and (c) Asphamin; (d–f) 590–555 cm\u003csup\u003e-1\u003c/sup\u003e range for mastics containing (d) neat bitumen, (e) Sasobit, and (f) Asphamin\u003c/p\u003e","description":"","filename":"image12.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/566846ef38e133a6d2c191d2.png"},{"id":79663937,"identity":"bff76f2b-3829-458c-989e-97c6e2797b70","added_by":"auto","created_at":"2025-04-01 09:57:16","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":348748,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of integral values for two spectral ranges across different mastic rejuvenations: (a–c) 650–610 cm\u003csup\u003e-1\u003c/sup\u003e range for mastics (a) unrejuvenated, (b) with Sylvaroad, and (c) with Storflux; (d–f) 810-785 cm\u003csup\u003e-1\u003c/sup\u003e range for mastics (d) unrejuvenated, (e) with Sylvaroad, and (f) with Storflux\u003c/p\u003e","description":"","filename":"image13.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/2e2a1ba4468f557785d22bf4.png"},{"id":79661374,"identity":"c08e19e9-a374-474e-8dae-bbc507b68d98","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":335062,"visible":true,"origin":"","legend":"\u003cp\u003eBox plots of integral values for two spectral ranges across different mastic rejuvenations: (a–c) 1300-1160 cm\u003csup\u003e-1\u003c/sup\u003e range for mastics (a) unrejuvenated, (b) with Sylvaroad, and (c) with Storflux; (d–f) 1720-1670 cm\u003csup\u003e-1\u003c/sup\u003e range for mastics (d) unrejuvenated, (e) with Sylvaroad, and (f) with Storflux\u003c/p\u003e","description":"","filename":"image14.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/90a8d960ae09c347f9128e5b.png"},{"id":79663936,"identity":"cdaebf3e-3e90-415a-9843-ea23a11ff388","added_by":"auto","created_at":"2025-04-01 09:57:16","extension":"png","order_by":15,"title":"Figure 15","display":"","copyAsset":false,"role":"figure","size":158570,"visible":true,"origin":"","legend":"\u003cp\u003eBox plot of integrals at 1770-1725 cm-1 range of mastics: (a) without warm mix additive, (b) with Sasobit, and (c) with Asphamin\u003c/p\u003e","description":"","filename":"image15.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/7692720413b8626f6ee5a89c.png"},{"id":79661404,"identity":"942f79dd-1ac3-4bfe-8c99-6bd6a73236f8","added_by":"auto","created_at":"2025-04-01 09:41:15","extension":"png","order_by":16,"title":"Figure 16","display":"","copyAsset":false,"role":"figure","size":148205,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Projection of normalized VIP scores before and after rejuvenation, (b) wavenumbers likely to remain irreversible after rejuvenation\u003c/p\u003e","description":"","filename":"image16.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/06bbc98b89d922c989e80ad4.png"},{"id":79664805,"identity":"2d00a6eb-f058-446c-960a-a5c2b1a32adf","added_by":"auto","created_at":"2025-04-01 10:05:15","extension":"png","order_by":17,"title":"Figure 17","display":"","copyAsset":false,"role":"figure","size":521734,"visible":true,"origin":"","legend":"\u003cp\u003eComparing the integral values of rejuvenated and unrejuvenated mastics at (a) 650-610 cm\u003csup\u003e-1\u003c/sup\u003e, (b) 785-770 cm\u003csup\u003e-1\u003c/sup\u003e, and (c) 1300-1160 cm\u003csup\u003e-1\u003c/sup\u003e\u003c/p\u003e","description":"","filename":"image17.png","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/979d313b7a53a041aa78ec15.png"},{"id":79666017,"identity":"0d0b2d75-b6e4-449a-82e3-56415b6a49a0","added_by":"auto","created_at":"2025-04-01 10:13:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6556120,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6342505/v1/964d94de-beba-4c10-a7e0-f70e2e068863.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eRejuvenation mechanisms in bituminous RAP mastics: insights from FTIR spectroscopy and multivariate discriminant analysis\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eThe increasing focus on sustainability and circularity in asphalt production and road construction has driven significant interest in incorporating higher proportions of reclaimed asphalt pavement (RAP) into new asphalt mixtures. RAP utilization presents an opportunity to reduce resource consumption and environmental impact, but it also introduces challenges due to the altered properties of the aged binder coating RAP (referred to as RAP binder). Aging mechanisms, such as oxidation and the loss of volatile components, lead to embrittlement and hardening of the binder, significantly impacting the performance of asphalt mixtures. Without adequate measures to address these changes, the incorporation of RAP can result in premature distresses such as cracking and fatigue cracking, thereby shortening the service life of asphalt structures [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe aging-induced changes in bitumen structure are commonly classified into reversible and irreversible categories. Oxidation, a major irreversible process, results in the formation of oxygen-containing functional groups such as sulfoxide (S\u0026thinsp;=\u0026thinsp;O) and carbonyl (C\u0026thinsp;=\u0026thinsp;O), as well as an increase in large molecular species due to dehydrogenation [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. While reversible changes can be counteracted through rejuvenation, irreversible changes remain permanent, even after applying recycling agents (RAs). These agents are designed to restore the colloidal balance of bitumen by rebalancing the asphaltene-to-maltene ratio [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], transitioning the binder structure from a \u0026ldquo;gel-type\u0026rdquo; to a \u0026ldquo;sol-gel type\u0026rdquo; system [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. However, excessive use of RAs can create a \"sol-type\" structure, making the binder more prone to deformation at high temperatures [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRAs introduce distinctive chemical components to restore bitumen's microstructure and rheological properties. Despite its effectiveness, complete restoration of aged binders\u0026rsquo; original characteristics is unattainable due to irreversible chemical changes. Therefore, optimizing RA content based on the RAP binders\u0026rsquo; aging level is critical to achieving the best restoration level without compromising durability. Several studies have evaluated rejuvenation efficiency using various experimental approaches, including mechanical, rheological, and microstructural assessments.\u003c/p\u003e \u003cp\u003eMechanical testing has demonstrated that RAs enhance the cracking resistance of RAP-containing asphalt mixtures while maintaining rutting resistance and moisture susceptibility [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Measures such as the penetration index have been proposed as indicators of RA effectiveness [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Rheological evaluations, involving dynamic shear rheometer (DSR) and bending beam rheometer (BBR) tests, have been used to assess the restoration of binder properties at both high and low temperatures [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. These studies have highlighted the ability of RAs to reduce the performance grade of aged binders in recycled mixtures.\u003c/p\u003e \u003cp\u003eMicroscopic techniques, such as atomic force microscopy (AFM), have investigated the morphological and microstructural changes in aged binders induced by RAs [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Significant differences have been reported in the effectiveness of various RAs, including vegetable oil, aromatic extracts, and bio-based agents, in altering the microstructure of aged binders [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Similarly, Fourier transform infrared (FTIR) spectroscopy and SARA fractionation have been utilized to characterize the chemical changes associated with rejuvenation. FTIR spectroscopy, in particular, has been used to monitor the absorption intensities of functional groups before and after rejuvenation, providing insights into the oxidative aging processes and the effectiveness of RAs [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. In this context, various methods have been proposed including the calculation of band areas as described in [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], employing an aging index tied on the carbonyl, sulfoxide, and aliphatic band [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], and measuring the carbonyl and sulfoxide indices per [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. However, questions remain regarding the reliability of FTIR indices for distinguishing between different RAs and their effectiveness due to their dependence on functional groups reflecting oxidative aging [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eGiven that the migration of aged components into the blended binder appears as chemical alterations, FTIR spectroscopy can be a suitable method for uncovering the mechanisms underlying the rejuvenation process. Using the extensive information provided by the FTIR spectroscopy, substantial information can be gained to characterize the changes induced by RJ incorporation and the blending of RAP binder with virgin bitumen. Beyond the commonly studied features and wavenumbers, such as carbonyl and sulfoxide function groups, chemometric analyses employing data science techniques\u0026mdash;specifically multivariate discriminant analysis\u0026mdash;have attracted attention for their potential to identify less prominent features (e.g., wavenumbers with lower absorption intensities) that play critical roles [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. This approach offers several advantages, including differentiating bitumen samples based on type and aging state [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], identifying bitumen sources and refining source [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], detecting polycyclic aromatic hydrocarbons (PAHs) in bitumen samples [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e], assessing aging evolution rates in bitumen mastics [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], and differentiating bitumen mastics based on the types of incorporated mineral fillers [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Overall, this analytical approach enables more accurate differentiation of bituminous samples according to specific criteria and provides insights into the underlying mechanisms driving these differences through feature engineering methodologies. This can be beneficial in recognizing the chemical changes induced in bitumen structure from a chemical point of view.\u003c/p\u003e \u003cp\u003eMultivariate discriminant analysis is a statistical approach that integrates data mining and machine learning techniques to classify and distinguish bituminous samples with predefined labels based on their chemical characteristics. This method identifies underlying patterns and relationships among wavenumbers that influence differentiation, considering variables such as aging level, RAP content, and additive type. The fundamental principle of multivariate discriminant analysis involves projecting data onto discriminant functions to enhance separation between groups. This process typically consists of two main steps: reducing the dataset\u0026rsquo;s dimensionality and classifying the reduced dataset according to predefined classes. By combining distinct algorithms, this hybrid technique aims to maximize the separation between predefined classes, facilitating dataset analysis and feature extraction.\u003c/p\u003e \u003cp\u003eVarious algorithms are employed for dimensionality reduction, including factor analysis (FA), principal component analysis (PCA), and partial least square regression (PLSR). The underlying principles and mathematical formulations of these algorithms are elaborated in [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. All dimensionality reduction methods focus on generating new features\u0026ndash;latent variables (LVs) or principal components (PCs)\u0026ndash;to simplify complex spectral data while retaining most of the variance. PCA identifies directions through PCs that capture the maximum variability in the dataset without considering labels, aiming solely to maximize variance. FA, on the other hand, seeks LVs that explain relationships within the dataset. Unlike these two methods, PLSR incorporates both predictors and responses, maximizing covariance between them to identify LVs that strengthen the relationship between independent and dependent variables. The primary objective of PLSR is to pinpoint features that contribute most to class separability.\u003c/p\u003e \u003cp\u003eThe effectiveness of these dimensionality reduction techniques in uncovering patterns within FTIR spectroscopy data of bituminous samples has been explored, proposing that the PLSR algorithm can show promises in generating LVs while providing variance importance in projection (VIP) scores, offering an intuitive approach to feature selection [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Following successful dimensionality reduction, linear discriminant analysis (LDA) has emerged as a robust classifier due to its inherent ability to generate characteristic components that project distinct classes onto a Cartesian coordination system.\u003c/p\u003e \u003cp\u003eIn the context of evaluating the effectiveness of rejuvenation, the mechanisms governing the performance of RJs remain insufficiently understood despite promising outcomes reported through experimental and mechanical investigations. To study the interactions between RJs and blended binders, this study aims to address these gaps by focusing on two critical aspects: (i) the extent to which the RAP binder fraction interacts with virgin bitumen to influence the properties of the blended binder, and (ii) the mechanisms by which distinct RJs alleviate the effects of aged components transferred from the RAP binder into the blended binder.\u003c/p\u003e \u003cp\u003eThe blending and rejuvenation phenomena in recycled asphalt mixtures will be examined using FTIR spectroscopy to address the research objectives and bridge existing knowledge gaps. The resulting data will be analyzed through multivariate discriminant analysis, specifically by employing PLSR coupled with LDA (PLSR-LDA), to uncover underlying patterns and principal mechanisms governing the interactions between RJs and aged components within RAP binders. The specific goals of this research are outlined as follows:\u003c/p\u003e\u003cp\u003e1- Investigating whether the blended binder in recycled mixtures contains aged and oxidized compounds in proportion to the recycling rate.\u003c/p\u003e \u003cp\u003e2- Determining the extent to which RJs can mitigate the presence of aged components and elucidating the mechanisms involved.\u003c/p\u003e \n\u003cp\u003e3- Identifying the types of aged components that can be retrieved and retained after rejuvenation.\u003c/p\u003e\n"},{"header":"2 Materials and methods","content":"\u003cp\u003eThe research aimed to address key objectives by preparing experimental samples at bitumen mastic scale. Fine RAP fractions, passing a 90 \u0026micro;m sieve size, were blended with basalt mineral fillers, virgin bitumen, two warm mix additives, and two commercial RJs. A constant filler-to-binder volumetric ratio of 27% was maintained to ensure consistent findings, in line with recommendations for dense-graded bituminous mixtures featuring a nominal maximum aggregate size of 19 mm [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSpecific gravity measurements were conducted for virgin filler, RAP filler, recovered RAP filler, and Asphamin filler, the latter being characterized by its filler form and moisture content. The base bitumen had a penetration grade of 50/70. Sasobit and Asphamin were incorporated at dosages of 3% and 0.3% by weight, respectively. RAP contents of 0%, 30%, 50%, 70%, and 100% were employed to examine a wide inclusion range. The RAP filler, sourced locally, had a binder content of 12.6%, determined using the ignition method per ASTM D6307 specifications.\u003c/p\u003e \u003cp\u003eEach mastic formulation was back-calculated based on the measured specific gravity of the components to achieve the target filler-to-binder volumetric ratio of 27%. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e outlines the 14 prepared mastic compositions, detailing the volumetric concentrations of RAP filler, virgin bitumen, RAP binder, virgin filler, and Asphamin, alongside the filler-to-binder volume ratios. The selected volumetric ratio led to mass ratios ranging from 1.1:1 to 1.18:1, aligning with dense-graded asphalt concrete compositions.\u003c/p\u003e \u003cp\u003eMastics were categorized based on their components: those incorporating only virgin fillers and/or warm mix additives were labeled \"VM\". In contrast, mastics containing RAP fillers represented varying levels of RAP replacement. For complete virgin filler replacement, virgin bitumen was used to maintain the target volumetric filler-to-binder ratio.\u003c/p\u003e \u003cp\u003eThe mastics were prepared using low-shear mixing at 160\u0026deg;C for unmodified samples and 135\u0026deg;C for those containing warm mix additives. Virgin filler was preheated at 160\u0026deg;C, while RAP fines were conditioned at 110\u0026deg;C for 2 hours [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. After component mixing, mechanical stirring was performed at 900 rpm. To prevent phase separation between bitumen and filler, the blends were poured into 4 mm thick metal molds.\u003c/p\u003e \u003cp\u003eFurthermore, two commercial rejuvenators (e.g., Storflux and Sylvaroad) were utilized. The optimal dosages for each were determined as outlined in Section 3 through a binder-fast-characterization-test (BTSV), according to European Standard EN 17643, conducted using a dynamic shear rheometer (DSR). Based on the determined optimal contents and the RAP binder amounts specified for each mastic variant in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, the required rejuvenator quantities were incorporated directly during the mastic preparation process.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComposition of Mastics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMastic ID\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMastic characteristics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eRejuvenation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eRAP filler volumetric concentration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"5\" nameend=\"c9\" namest=\"c5\"\u003e \u003cp\u003eVolumetric percentages in bituminous mastic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eFiller-to-binder volumetric ratio\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eVirgin binder\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRAP binder\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eVirgin filler\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eRecovered RAP filler\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAsphamin\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVirgin mastic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"14\" rowspan=\"15\"\u003e \u003cp\u003eUnrejuvenated; Storflux and Sylvaroad RJs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e72.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e27.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\" morerows=\"14\" rowspan=\"15\"\u003e \u003cp\u003e0.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSaso-VM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVirgin mastic with Sasobit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e72.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e27.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsph-VM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVirgin mastic with Asphamin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e72.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e22.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e5.38\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRAP-30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 30% RAP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e69.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e18.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e8.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRAP-50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 50% RAP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e66.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e13.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e13.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRAP-70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 70% RAP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e64.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e8.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e18.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRAP-100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 100% RAP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e60.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e27.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSaso-RAP-30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 30% RAP and Sasobit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e69.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e18.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e8.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSaso-RAP-50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 50% RAP and Sasobit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e66.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e13.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e13.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSaso-RAP-70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 70% RAP and Sasobit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e64.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e8.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e18.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSaso-RAP-100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 100% RAP and Sasobit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e60.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e27.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsph-RAP-30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 30% RAP and Asphamin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e68.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e14.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e8.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e5.40\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsph-RAP-50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 50% RAP and Asphamin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e65.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e8.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e13.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e5.40\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsph-RAP-70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 70% RAP and Asphamin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e63.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e2.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e19.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e5.40\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsph-RAP-100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMastic with 100% RAP and Asphamin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e61.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e10.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e22.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e5.42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe mastics underwent chemical measurement using FTIR spectroscopy with a Bruker Alpha device equipped with an attenuated total reflection (ATR) unit. Spectral intensities were recorded over wavenumbers ranging from 4000 to 400 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e with 24 scans for each spectrum. To ensure reproducibility, five samples of each testing variant were prepared, each measured five times, according to [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e"},{"header":"3 Optimal RJ content","content":"\u003cp\u003eTo determine the appropriate dosage of rejuvenators for restoring the desired properties of virgin bitumen, BTSV testing was utilized to assess the rheological characteristics of bitumen samples according to [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. This method involves measuring bitumen's complex modulus and phase angle over a continuous temperature increase. Key parameters derived from this test include: (i) equi-shear modulus temperature at which the complex modulus equals 15 kPa (T\u003csub\u003eBTSV\u003c/sub\u003e) and (ii) the corresponding phase angle (δ\u003csub\u003eBTSV\u003c/sub\u003e). According to methodologies outlined in [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], the examined bitumen samples can be differentiated into classes reflecting their penetration grades and modification level.\u003c/p\u003e \u003cp\u003eThis investigation performed BTSV testing on virgin bitumen (50/70 penetration grade) and binder recovered from RAP, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea,b. The results showed that the virgin binder showed a T\u003csub\u003eBTSV\u003c/sub\u003e of 51.53 ℃ and a δ\u003csub\u003eBTSV\u003c/sub\u003e of 78.89\u0026deg;. On the other hand, the RAP binder showed a T\u003csub\u003eBTSV\u003c/sub\u003e of 68.61 ℃ and a δ\u003csub\u003eBTSV\u003c/sub\u003e of 76.91\u0026deg;, displaying the effect of aging, including increased stiffness and embrittlement. These findings, mapped according to the classification framework detailed in [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe results of the BTSV testing indicated that the binder characteristics were positioned, with the combination of T\u003csub\u003eBTSV\u003c/sub\u003e and δ\u003csub\u003eBTSV\u003c/sub\u003e correctly reflecting the penetration grade of the bitumen. In contrast, the RAP binder characteristics, represented by the red dot, showed a significant increase in binder stiffness due to aging, positioning the sample closer to the category of bitumen with a 20/30 penetration grade. To counteract the aging effects, various contents of Sylvaroad and Storflux were blended with RAP binder, and the resultant blends were subjected to BTSV testing. The projections from the BTSV testing revealed that employing 10% w/w Sylvaroad and 15% w/w Storflux significantly restored the desired characteristics of the RAP binder, although failed to return to the 50/70 penetration grade category fully.\u003c/p\u003e \u003cp\u003eThe data in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e show that increasing the RJ contents reduced the T\u003csub\u003eBTSV\u003c/sub\u003e while only slightly restored δ\u003csub\u003eBTSV\u003c/sub\u003e, thereby improving the penetration grades. Based on the changes depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, the optimal RJ contents were determined for the preparation of rejuvenated RAP mastics, as outlined in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The testing results for the RAP binder with 15% w/w Storflux and 10% w/w Sylvaroad are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec,d.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"4 Exploration of chemical markers","content":"\n\u003ch3\u003e4-1- Formal analyses\u003c/h3\u003e\n\u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea-c illustrates a sample of FTIR spectroscopy from RAP mastics incorporating various RJs. Relying on their standard deviations, the region from 1600 to 400 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e reflects the majority of the changes due to variation in RAP concentrations, close to the bitumen fingerprint region [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Only this critical region is shown for brevity, although all recorded wavenumbers will be considered during the classification and pattern recognition tasks. In addition, the spectroscopical absorption of RJs, mineral filler resources, and Asphamin are depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed-f.\u003c/p\u003e \u003cp\u003eThe spectral data in this figure shows that higher RAP content enhanced absorption intensity within 600\u0026thinsp;\u0026minus;\u0026thinsp;400 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 1700\u0026thinsp;\u0026minus;\u0026thinsp;1150 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e ranges while reducing it in 1150\u0026thinsp;\u0026minus;\u0026thinsp;800 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. Comparing these ranges, the absorption peaks within 1000\u0026thinsp;\u0026minus;\u0026thinsp;400 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e are likely attributed to changes in the mineral composition of bituminous mastics, whereas wavenumbers from 1700\u0026thinsp;\u0026minus;\u0026thinsp;1400 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e reflect the chemical changes in the carbonyl, amides, and aromatic bands. Notably, the sulfoxide band, typically around 1030 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, may be obscured by the mineral fillers' interface. Additionally, the presence of RJs in RAP mastics influences absorption intensities, which are not visible in this figure.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed shows Sylvaroad with a prominent peak at 1739 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, indicating high ester content. Esters are recognized for their ability to interact with oxidized compounds in aged bitumen, reducing their concentration. The 1370 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e peak in Storflux indicates symmetric bending of methyl (CH₃) groups, highlighting a higher presence of aliphatic components in Storflux compared to Sylvaroad. Aliphatic hydrocarbons in rejuvenators help restore hydrocarbon loss from bitumen aging, balancing polar and non-polar components in the bitumen structure.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo identify outliers within the dataset, a modified Z-score test, equipped with Median Absolute Deviation (MAD) as outlined in [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e], was conducted using Eq.\u0026nbsp;\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. In which, x\u003csub\u003ei\u003c/sub\u003e represents the data point (absorbance at each wavenumber) and the threshold of |Z\u003csub\u003emodified\u003c/sub\u003e| \u0026gt; 3.0 is controlled to detect anomalies. With five repetitions performed for each sample, the modified Z-score assessed the variability within the measurements:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:{Z}_{modified}=\\frac{0.6745\\times\\:({X}_{i}-Median\\:of\\:all\\:spectra)}{Median\\left(\\left|{X}_{i}-Median\\right|\\right)}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Z-scores Box plots of FTIR spectroscopy measurements to identify repetitions outliers: (a) neat bitumen-unrejuvenated, (b) neat bitumen-Sylvaroad, (c) neat bitumen-Storflux, (d) Sasobit modified bitumen-unrejuvenated, (e) Sasobit modified bitumen-Sylvaroad, (f) Sasobit modified bitumen-Storflux, (g) Asphamin modified bitumen-unrejuvenated, (h) Asphamin modified bitumen-Sylvaroad, and (i) Asphamin modified bitumen-Storflux\u003c/p\u003e\n\u003ch3\u003e4-2- Chemical markers reflecting the presence of RJs\u003c/h3\u003e\n\u003cp\u003eThe dataset was initially analyzed to distinguish between the samples according to RJ types, resulting in a classification involving labels of control mastic, unrejuvenated, Storflux, and Silvaroad. The dimensionality of the dataset was reduced by applying the algorithm detailed in the reference [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. The transformation of the dataset into a new coordinate system generated 17 LVs, determined by plotting a Scree plot illustrating the cumulative variance explained versus the number of LVs. Processing the LVs\u0026mdash;which were generated based on different derivation levels\u0026mdash;using LDA classified the dataset into distinct classes, as presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. It is worth recalling that the present research adopted a multilevel approach to conduct multivariate discriminant analysis to identify influential chemical components while evaluating strong, prominent, and subtle peaks.\u003c/p\u003e \u003cp\u003eFor classification purposes, 5-fold cross-validation was employed, yielding a mean accuracy of 0.94, with a precision of 0.97, recall of 0.96, and F1-score of 0.96. In this context, precision reflects the proportion of correctly labeled samples among those predicted by the model, recall indicates the model's ability to identify all relevant instances in each class, and the F1-score, as a harmonic mean of precision and recall, reflects the model's effectiveness in both identifying the population of each class and labeling them correctly. The only failure that the model faced in identifying the labels occurred in distinguishing Storflux and Sylvaroad.\u003c/p\u003e \u003cp\u003eClassification based on the first two LDA components (capturing over 90% of system variability) showed similar chemical characteristics between the two rejuvenated mastic classes and the control. However, classification using LDA components 1 and 3 positioned control and rejuvenated mastics near zero for the 3rd LDA component, while distinguishing between the two RJ types. This distinction was controlled by wavenumbers with smaller normalized VIP scores.\u003c/p\u003e \u003cp\u003eThroughout this classification, normalized VIP scores were calculated and visualized in Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef, and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ei, highlighting influential wavenumbers across derivation levels. Interestingly, the C\u0026thinsp;=\u0026thinsp;O stretching vibration of carbonyl functional groups at ~\u0026thinsp;1745 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e exhibited the highest normalized VIP scores.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo identify key spectral regions governing multilevel classification, the highlighted wavenumbers in Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef, and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ei were projected onto a sample spectrum in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea and subsequent subfigures detailing their corresponding regions. This figure highlights two key observations: (i) in Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec\u0026ndash;d, higher RAP contents significantly altered absorption intensities; (ii) Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee reveals changes in the Sylvaroad-associated absorption peak, where higher RAP content\u0026mdash;and subsequently higher RJ content\u0026mdash;intensified absorption, which can be inferred as residue RJs due to partial blending of RAP binder and RJ.\u003c/p\u003e \u003cp\u003eUsing OPUS software, integrals were calculated for the spectral ranges 590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, 1330\u0026thinsp;\u0026minus;\u0026thinsp;1275 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, 1370\u0026thinsp;\u0026minus;\u0026thinsp;1360 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, and 1770\u0026thinsp;\u0026minus;\u0026thinsp;1725 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. Beyond their chemical significance and the trends shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e, two key aspects were considered for feature selection: (i) the contribution of these bands to the classification in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and (ii) their sensitivity to variables such as RAP content, RJs, and warm mix additives.\u003c/p\u003e \u003cp\u003eTo assess feature importance, four ML algorithms\u0026mdash;random forest, logistic regression, XGBoost, and CatBoost\u0026mdash;were applied to classify the dataset based on the integrals. The results consistently ranked the absorption bands in order of importance:1758\u0026thinsp;\u0026minus;\u0026thinsp;1715 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026gt; 1370\u0026thinsp;\u0026minus;\u0026thinsp;1360 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026gt; 1330\u0026thinsp;\u0026minus;\u0026thinsp;1275 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e \u0026gt; 590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. The dependency of these spectral ranges on the testing variables was further analyzed using the same ML algorithms. The results indicated that the integral values of 1758\u0026thinsp;\u0026minus;\u0026thinsp;1715 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 1370\u0026thinsp;\u0026minus;\u0026thinsp;1360 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e were primarily influenced by RJs rather than RAP content or warm mix additives. In contrast, the integral values of 1330\u0026thinsp;\u0026minus;\u0026thinsp;1275 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e were more dependent on RAP content. The analysis also suggested the reflection of warm mix additives within these ranges.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFrom a chemical perspective, the 1758\u0026thinsp;\u0026minus;\u0026thinsp;1715 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range corresponds to carbonyl stretching in carboxylic acids (-COOH) [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e] and C\u0026thinsp;=\u0026thinsp;O stretching in esters [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], both of which form during aging. In rejuvenated bituminous mixtures, esters in this range may indicate resistance to aging and the effectiveness of rejuvenation [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. The spectral results of Sylvaroad, shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed, exhibit a significant peak within this range, suggesting that the prominent peaks in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ed correspond to the presence of this specific RJ.\u003c/p\u003e \u003cp\u003eThe peaks at 1370\u0026thinsp;\u0026minus;\u0026thinsp;1360 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e correspond to C-H bending vibrations in hydrocarbons, specifically in methyl (CH\u003csub\u003e3\u003c/sub\u003e) and methylene (CH\u003csub\u003e2\u003c/sub\u003e) groups. Methyl groups are primarily found in saturated and aromatic hydrocarbons, while methylene groups are more common in aliphatic hydrocarbons. This range may indicate the influence of recycling agents, where increased intensity suggests a potential restoration of aliphatic components in bitumen, enhancing flexibility\u0026mdash;consistent with the feature importance findings. A comparison of the integrals in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb highlights the distinction between Storflux-modified samples and others. Additionally, the absorption intensity of various RJs in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed exhibits a characteristic peak at 1370 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, confirming differences in their integral values.\u003c/p\u003e \u003cp\u003eThe 1330\u0026thinsp;\u0026minus;\u0026thinsp;1275 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range corresponds to C-O stretching vibrations in hydroxyl (-OH) groups, commonly found in alcohols or phenolic compounds [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. The oxidation of bitumen in RAP binders likely increases hydroxyl group concentrations as RAP content rises. This aligns with the feature importance results, where RAP content was identified as a key factor influencing this range (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb). However, Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea indicates minimal changes in integral values after RJ incorporation, suggesting these oxidative changes remain largely irreversible, consistent with findings in the literature. Since this range does not significantly differentiate mastics before and after rejuvenation, it primarily distinguishes control mastics from RAP mastics.\u003c/p\u003e \u003cp\u003eThe absorption at 590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e in the zero-order dataset reflects metal-oxygen bending, attributed to phosphate (PO\u003csub\u003e4\u003c/sub\u003e\u003csup\u003e3\u0026minus;\u003c/sup\u003e) compounds, as indicated by Table\u0026nbsp;7.5 in [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. The increasing intensity in this range with higher RAP content suggests a notable contribution from phosphate compounds, likely originating from RAP fillers. A comparison of FTIR spectra for RAP and basalt fillers (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee) further supports this, with a pronounced absorption near 500 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e reinforcing the presence of phosphate-related metal oxides introduced by RAP fillers into the mastic.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003e4-3- Chemical markers reflecting the effectiveness of Sylvaroad\u003c/h3\u003e\n\u003cp\u003eUpon distinguishing between various RJs in RAP mastics, their mechanisms in mitigating chemical changes induced by RAP binders remain unclear. This section identifies chemical markers reflecting Sylvaroad\u0026rsquo;s effectiveness in mitigating RAP binder-induced changes. Key markers distinguishing RAP mastics before and after rejuvenation are examined, where those losing significance post-rejuvenation indicate Sylvaroad\u0026rsquo;s influence.\u003c/p\u003e \u003cp\u003eFigures \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea, \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ee, and \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ei show classification before rejuvenation across different derivation orders, while Figs.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ec, \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eg and \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ek depict Sylvaroad-rejuvenated mastics. Normalized VIP scores from zero-order datasets suggest that Syvlaroad\u0026rsquo;s effect is not strongly reflected in major absorption peaks, except at 2786 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, linked to C\u0026ndash;H stretching in aliphatic hydrocarbons, possibly indicating a subtle oxidation effect. In the first derivative dataset, peaks at 567 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, 803 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, and 1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e appear before and after rejuvenation, but Sylvaroad notably reduces the significance of 2748 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. The second derivative dataset does not clearly distinguish Sylvaroad\u0026rsquo;s impact based on normalized VIP scores.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo further examine these wavenumbers, they are projected onto a spectral sample representing the effect of increasing RAP concentrations. This projection, shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e, highlights absorption regions capturing the influence of RAP variation and, presumably, Sylvaroad. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e summarizes these ranges, detailing their derivation level, corresponding functional groups, and interpretation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e suggests that the presence of RAP fractions in both control and Sylvaroad-rejuvenated mastics can be assessed by evaluating the integrals of specific absorption ranges. The 489\u0026thinsp;\u0026minus;\u0026thinsp;446 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range (~\u0026thinsp;Figure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eb), associated with the Si-O band, can be a marker of minerals such as fillers or organic compounds in Asphamin. The 590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range (~\u0026thinsp;Figure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ec) corresponds to phosphate (PO\u003csub\u003e4\u003c/sub\u003e\u003csup\u003e3\u0026minus;\u003c/sup\u003e) compounds, indicating potential mineral compounds. The 650\u0026thinsp;\u0026minus;\u0026thinsp;610 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range (~\u0026thinsp;Figure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ed), linked to aromatic out-of-plane vibrations of hydrocarbons or sulfates, and the 810\u0026thinsp;\u0026minus;\u0026thinsp;785 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range (~\u0026thinsp;Figure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ee), associated with out-of-plane C-H bending vibrations, particularly in meta-distributed benzene rings per [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e], are also key indicators.\u003c/p\u003e \u003cp\u003eTypically, aging reduces absorption intensity in the 810\u0026thinsp;\u0026minus;\u0026thinsp;785 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e], Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ee, as seen in the comparison between RAP and virgin binders in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed. However, the observed peak may also originate from recovered RAP fillers, with peaks at 770 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, 789 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, and 879 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee. Specifically, the 789 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e peak indicates the quartzite nature of RAP fillers, while the 879 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e peak points to the carbonate [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. These findings suggest that increasing RAP content creates a local peak in this range, useful for distinguishing RAP contents based on mineral components.\u003c/p\u003e \u003cp\u003eThe 1300\u0026thinsp;\u0026minus;\u0026thinsp;1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range (~\u0026thinsp;Figure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ef) likely reflects various chemical compounds, including S\u0026thinsp;=\u0026thinsp;O stretching in sulfonates, sulfates, and sulfoxides, as well as C-O stretching in ethers and carboxylic acid derivatives [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Mirwald et al. [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e] and Primerano et al. [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] attributed the increased absorption in this range to the overall increase in bitumen\u0026rsquo;s polarity due to oxidative aging. This increase in absorption, correlates with higher RAP content, indicating the migration of oxidized components from the RAP binder to the blended binder, consistent with findings in [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. The upward trend in absorption suggests a higher concentration of oxidized compounds in RAP mastics.\u003c/p\u003e \u003cp\u003eFinally, the 1770\u0026thinsp;\u0026minus;\u0026thinsp;1725 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range (~\u0026thinsp;Figure \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eg) represents C\u0026thinsp;=\u0026thinsp;O stretching, a known marker of oxidative aging, even may also be attributed to Sylvaroad, which shows a strong absorption at 1739 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSignificant wavenumbers based on classification are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProposed range for integration\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eZero-order spectrum\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1st derivative\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2nd derivative\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eFunctional group\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eInterpretation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e489\u0026thinsp;\u0026minus;\u0026thinsp;446 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSi\u0026ndash;O bending or Al\u0026ndash;O vibrations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMineral filler contribution (e.g., aggregates, additives like Asphamin)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMetal-oxygen bending\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePhosphate (PO\u003csub\u003e4\u003c/sub\u003e\u003csup\u003e3\u0026minus;\u003c/sup\u003e) compounds from RAP fillers\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e650\u0026thinsp;\u0026minus;\u0026thinsp;610 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAromatic C\u0026ndash;H out-of-plane bending\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAromatic ring structures from aged bitumen in RAP\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e810\u0026thinsp;\u0026minus;\u0026thinsp;785 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAromatic C\u0026ndash;H out-of-plane bending\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAged bitumen\u0026rsquo;s aromatic character, possibly modified by rejuvenators\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1300\u0026thinsp;\u0026minus;\u0026thinsp;1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eC\u0026ndash;O stretching\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFormation of oxygenated compounds such as alcohols and esters\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1720\u0026thinsp;\u0026minus;\u0026thinsp;1670 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eC\u0026thinsp;=\u0026thinsp;O stretching (carbonyls, ketones, esters)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eOxidation products presumably from RAP binder aging\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1770\u0026thinsp;\u0026minus;\u0026thinsp;1725 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✘\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCarbonyl groups such as esters, aldehydes, and ketones\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eThe marker of Sylvaroad\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003e4-4- Chemical markers reflecting the effectiveness of Storflux\u003c/h3\u003e\n\u003cp\u003eSimilarly, the subset of RAP mastics incorporating Storflux was analyzed using the same approach, leading to a multilevel classification based on different derivation levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e). Notably, this classification revealed that no significant pattern emerged from the zero-order dataset, where the distinction shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ea primarily captured noise, resulting in inconclusive findings. Evaluation of the first derivative results, Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eb, identified wavenumbers at 656 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, 1729 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, and 3249 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e as significant. The second derivative analysis, Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ec, highlighted wavenumbers at 658 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 3077 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e as important. These frequencies suggest that the distinction between mastic classes is primarily governed by carbonyl compounds, aromatic ring structures, and O\u0026ndash;H stretching vibrations. The normalized VIP scores corresponding to each classification are presented alongside the classification results in Figs.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eb, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ed, and \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ef.\u003c/p\u003e \u003cp\u003eProjecting these wavenumbers onto a spectral sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e) revealed two key absorption regions likely contributing to the observed distinctions: 650\u0026thinsp;\u0026minus;\u0026thinsp;610 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 1670\u0026ndash;1720 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. The former region was identified from both the first and second derivatives, while the latter was distinguished through first-derivative analysis.\u003c/p\u003e \u003cp\u003eThe 650\u0026ndash;610 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range can be attributed to aromatic out-of-plane bending vibrations in hydrocarbons and S-O bending vibrations in inorganic sulfates, as indicated by Tables\u0026nbsp;2.6, 6.5, and 7.1 in [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. The reduction in intensity absorption within this range, due to increased RAP content, aligns with the spectral results of RAP fillers shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee. The absorption within the 1670\u0026ndash;1720 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range captures the accumulation of carbonyl compounds due to increased RAP concentrations.\u003c/p\u003e \u003cp\u003eThe prominence of these two absorption regions, independent of the wavenumbers identified in unrejuvenated mastics, suggests that they serve as chemical markers correlated with oxidation and RAP mineral fillers, reflecting higher RAP contents. Furthermore, these findings indicate that Storflux played a role in restoring the significant maltene loss associated with aging.\u003c/p\u003e \u003cp\u003eAlthough higher RAP contents are typically identified by an increase in sulfoxide and/or carbonyl compounds, the interaction between basalt filler and the accumulation of Si-containing compounds likely masked oxidation markers within the sulfoxide region for the mastics tested in this study. Instead, evaluating the carbonyl functional groups, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003ec, indicates that higher RAP contents lead to the accumulation of oxidative compounds.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003e4-5- Pattern recognition\u003c/h3\u003e\n\u003cp\u003eMultivariate discriminant analysis identified several absorption ranges that distinguish mastic classes with varying RAP concentrations before and after rejuvenation. The analysis revealed the formation of distinctive differentiation patterns influenced by Sylvaroad and Storflux. Notably, the incorporation of Storflux, unlike Sylvaroad, significantly altered the differentiation pattern, as reflected in the normalized VIP scores computed at various derivation levels.\u003c/p\u003e \u003cp\u003eThis section aims to refine the observed patterns and elucidate the nonlinear relationships shaping these classifications. To achieve this, spectral datasets underwent integration within the identified ranges using OPUS software. Among these, the 489\u0026thinsp;\u0026minus;\u0026thinsp;446 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e regions could potentially serve as mineral composition markers within the blended mastics. Since each mixture contained three distinct mineral sources\u0026mdash;virgin filler, RAP mineral particles, and Asphamin\u0026mdash;the effect of increasing RAP content was evaluated separately through the corresponding integral values, as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eTwo key criteria were considered in this evaluation: (1) changes in integral values due to increasing RAP concentrations and (2) the distribution of recorded data points within each box, representing the effects of three RJ inclusion levels (unrejuvenated, Sylvaroad, and Storflux). The data distribution analysis indicates that RJ inclusion had a negligible effect on the integral values of the 590\u0026ndash;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e region and a relatively minor influence on the 489\u0026thinsp;\u0026minus;\u0026thinsp;446 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range. Consequently, these two regions appear to primarily capture the inclusion of higher RAP contents, serving as chemical markers of mineral composition.\u003c/p\u003e \u003cp\u003eFurther assessment of RAP content variations revealed a progressive increase in integral values for mastics containing neat bitumen and Sasobit, whereas an inverse trend was observed in mastics with Asphamin. Given that Asphamin exhibits a pronounced peak at 465 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, it contributes to a notable increase in absorption intensity and integral values. However, as RAP content increases, the intensity difference between Asphamin mastics and other variants diminishes. This suggests that Asphamin introduces Al\u0026ndash;O absorption, likely from zeolite, while RAP contributes to Si\u0026ndash;O absorption. The interaction between these two mineral phases at higher RAP concentrations likely leads to spectral flattening, reducing overall integral values. This phenomenon may indicate a redistribution effect, wherein Al\u0026ndash;O and Si\u0026ndash;O vibrations disperse, leading to a more uniform spectral response\u0026ndash;an observation that requires further validation.\u003c/p\u003e \u003cp\u003eRegarding the 590\u0026thinsp;\u0026minus;\u0026thinsp;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e integral values, an increase in RAP content resulted in a gradual decline, regardless of the type of warm mix additives used. The lower absorption intensities in this range, compared to those of basalt fillers, suggest a masking effect on sulfate-related absorptions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo evaluate changes in the composition of aromatic C\u0026ndash;H out-of-plane bending, observed in the 650\u0026thinsp;\u0026minus;\u0026thinsp;610 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 810\u0026thinsp;\u0026minus;\u0026thinsp;785 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e regions, integral values were plotted in Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e. The results indicate a continuous decline in absorbance intensities within the 650\u0026thinsp;\u0026minus;\u0026thinsp;610 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range with increasing RAP concentrations. A comparison of subfigures corresponding to different rejuvenation levels suggests a slight but noticeable influence of rejuvenation; however, the overall trend remained unchanged. Conversely, an increase in RAP concentration led to a gradual rise in integral values within the 810\u0026thinsp;\u0026minus;\u0026thinsp;785 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range, though at a slower rate. These trends are chemically consistent with the interpretation of the associated spectral regions, as discussed earlier.\u003c/p\u003e \u003cp\u003eOn the other hand, this figure highlights two key observations: (i) the variability introduced by warm mix additives, which is reflected in the scatter of recorded data points within each box and indicates their influence within each mastic class; and (ii) the negligible contribution of RJs in mitigating aging-related compounds that migrated from the RAP binder into the blended binder. The latter point suggests that the spectral features identified correspond to aromatic C\u0026ndash;H out-of-plane bending primarily capturing oxidation-induced chemical changes that remain irreversible despite rejuvenation efforts.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo assess the regions indicative of oxygen-containing markers, the integral values corresponding to the 1300\u0026thinsp;\u0026minus;\u0026thinsp;1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 1720\u0026thinsp;\u0026minus;\u0026thinsp;1670 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e regions were plotted in Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e14\u003c/span\u003e. The 1300\u0026thinsp;\u0026minus;\u0026thinsp;1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range, associated with the formation of oxygenated compounds such as esters and alcohols, exhibited a continuous increase in integral values, irrespective of the rejuvenation technique employed. However, a detailed comparison of data at the same RAP concentration level revealed that Sylvaroad led to a slight reduction in integral values, particularly at higher RAP concentrations.\u003c/p\u003e \u003cp\u003eSimilarly, the integral values within 1720\u0026thinsp;\u0026minus;\u0026thinsp;1670 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, which reflect the formation of carbonyl groups, showed a gradual increase, although at a slower rate. This trend suggests the migration of oxygenated compounds from the RAP binder into the blended binder, with Sylvaroad exerting a rather moderate mitigating effect on this transformation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe absorption region at 1770\u0026thinsp;\u0026minus;\u0026thinsp;1725 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e has previously been attributed to the characteristic absorption peak of Sylvaroad (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed). The computed integral values, as presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e15\u003c/span\u003e, indicate a distinct separation of mastics containing Sylvaroad from the others, regardless of the type of warm mix additives used.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003e4-6- Irreversible components post-rejuvenation\u003c/h3\u003e\n\u003cp\u003ePrevious sections examined chemical components associated with RJs and those reflecting varying RAP concentrations after rejuvenation. However, a key question remains: which aging-induced chemical changes persist in post rejuvenation? To address this, the feature extraction results, based on PLSR-LDA approach, were employed to identify chemical markers that distinguish mastic classes both before and after rejuvenation. This approach focuses on detecting features that were significant in unrejuvenated mastics and remain relevant even after rejuvenation, highlighting irreversible chemical changes.\u003c/p\u003e \u003cp\u003eInitially, normalized VIP scores, based on the first derivative dataset, were obtained from analyses of mastics before rejuvenation and mapped against those post-rejuvenation\u0026mdash;in which the post-rejuvenated class involved both RJs. Figure\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e16\u003c/span\u003ea presents the projection of normalized VIP scores from post-rejuvenation classification onto the pre-rejuvenation dataset.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e16\u003c/span\u003eb provides an enlarged view of wavenumbers with higher normalized VIP scores in both pre- and post-rejuvenation analyses. This indicates that the selected wavenumber range retains valuable information for distinguishing between mastic classes in both conditions. In other words, these ranges play a critical role in differentiating mastics with varying RAP percentages, regardless of rejuvenation. Therefore, these wavenumber ranges may represent irreversible chemical markers that persist after rejuvenation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe highlighted wavenumbers and their associated absorption ranges in Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e16\u003c/span\u003eb closely align with those identified in previous sections. As previously discussed, the wavenumbers at 642 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 768 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e are linked to mineralogical changes in RAP mastics, while 1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e is associated with oxidation phenomena.\u003c/p\u003e \u003cp\u003eFor unrejuvenated RAP mastics, similar absorption patterns were observed as in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, though with varying intensity levels, which are omitted here for brevity. To assess the correlation and consistency of absorption intensities across these ranges, integral values before and after rejuvenation were compared (Fig.\u0026nbsp;\u003cspan refid=\"Fig17\" class=\"InternalRef\"\u003e17\u003c/span\u003e). This figure presents the average integral values for the respective ranges alongside their standard deviations, evaluating the impact of RJs on these absorption regions. Two key observations can be drawn from this figure: (i) variations in standard deviations and the (ii) minimal impact of rejuvenation on absorption intensity.\u003c/p\u003e \u003cp\u003eThe data illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig17\" class=\"InternalRef\"\u003e17\u003c/span\u003e further validates the reliability of the PLSR-LDA model. As anticipated, the impact of rejuvenation on absorption regions related to the mineralogical characteristics of mastics was minimal. However, Fig.\u0026nbsp;\u003cspan refid=\"Fig17\" class=\"InternalRef\"\u003e17\u003c/span\u003ea reveals relatively high standard deviations for integral values, likely due to the influence of warm mix additives, particularly Asphamin, which may alter the mineral composition of mastics by introducing metals and mineral oxides. In contrast, Fig.\u0026nbsp;\u003cspan refid=\"Fig17\" class=\"InternalRef\"\u003e17\u003c/span\u003ec indicates that the absorption range of 1300\u0026ndash;1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e shows negligible variation with rejuvenation, accompanied by small standard deviation values. This range can therefore be identified as indicative of irreversible changes post-rejuvenation.\u003c/p\u003e \u003cp\u003eThe magnitude of the integrals for these wavenumber ranges also provides a basis for distinguishing mastics based on RAP content, both before and after rejuvenation. However, other regions and functional groups may also influence the blending of RAP binder with virgin bitumen. These effects potentially masked by the more prominent influence of mineral components. To overcome this limitation, further evaluations at the binder scale\u0026mdash;by isolating the blended binder from recycled mixtures and excluding mineral fillers\u0026mdash;could offer valuable insights. These investigations may provide a more refined understanding of irreversible chemical markers post-rejuvenation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"5 Conclusions and recommendations","content":"\u003cp\u003eThis study explored the rejuvenation mechanisms in bituminous RAP mastics using FTIR spectroscopy and multivariate discriminant analysis. The\u0026nbsp;goal\u0026nbsp;was to examine the chemical between RAP binder, virgin bitumen, warm mix additives, and RJs to uncover the the underlying mechanisms of RAP binder rejuvenation. To avoid analytical biases from inconsistencies in the filler-to-binder ratio, mastics with varying RAP concentrations were produced using a precise mix design,\u0026nbsp;maintaining a constant volumetric value. BTSV testing was conducted to identify the optimal RJ content, followed by a hybrid PLSR-LDA analysis of FTIR results, leading to the following findings:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eThe multilevel approach employed in this study, combining PLSR-LDA with normalized VIP scores, successfully identified key FTIR frequencies responsible for differentiating RAP mastics across various derivative levels.\u003c/li\u003e\n \u003cli\u003eClassification and feature extraction tasks to explore the presence of RJs in RAP mastics revealed the dominant role of absorption bands at 1330-1275 cm-1 (~C-O stretching vibrations in hydroxyl groups), 1370-1360 cm-1 (~C-H bending vibrations in hydrocarbons), and 1758-1715 cm-1 (~carbonyl stretching in carboxylic acids and esters). Among these ranges, the first is attributed to oxidative chemical changes, distinguishing control mastics, unrejuvenated RAP mastics, and rejuvenated mastics, reflecting irreversible chemical changes. The latter two ranges correspond to specific chemical markers associated with the RJs used.\u003c/li\u003e\n \u003cli\u003eInvestigation of the RJs impacts, three categories of chemical markers were identified as reflecting the changes post-rejuvenation: ranges linked to mineralogical characteristics (i.e., 489-446 cm\u003csup\u003e-1\u003c/sup\u003e and 590-555 cm\u003csup\u003e-1\u003c/sup\u003e), ranges attributed to aromatic markers (i.e., 650-610 cm\u003csup\u003e-1\u003c/sup\u003e and 810-785 cm\u003csup\u003e-1\u003c/sup\u003e), and ranges reflecting the changes within the oxygen containing frequencies (i.e., 1300-1160 cm\u003csup\u003e-1\u003c/sup\u003e and 1720-1670 cm\u003csup\u003e-1\u003c/sup\u003e). This observation suggests that varying RAP contents in mastics post-rejuvenation were identified based on both fillers\u0026rsquo; mineralogical characteristics and aging-induced changes.\u003c/li\u003e\n \u003cli\u003eTo identify aging-induced chemical markers that remain irreversible after rejuvenation, three absorption ranges were identified: 650-610 cm\u003csup\u003e-1\u003c/sup\u003e, 785-770 cm\u003csup\u003e-1\u003c/sup\u003e, and 1300-1160 cm\u003csup\u003e-1\u003c/sup\u003e. Evaluating their integral values before and after RJ incorporation showed consistent values, indicating the critical role of these regions in distinguishing mastic classes and their stability post-rejuvenation. Among these, the 1300-1160 cm\u003csup\u003e-1\u003c/sup\u003e range is linked to oxidation, reflecting irreversible chemical changes post-rejuvenation.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgment\u003c/h2\u003e \u003cp\u003eThe authors wish to extend their appreciation to the Alexander von Humboldt Foundation for their support of this project through the Georg Forster Postdoctoral Fellowship awarded to the first author.\u003c/p\u003e\u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eData will be made available on request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMocelin DM, Isied MM, da Costa RF, Castorena C (2024) Availability adjusted mix design method as a tool to mitigate the adverse effects of RAP on the performance of asphalt mixtures. Constr Build Mater 422. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2024.135813\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2024.135813\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMotevalizadeh SM, Mollenhauer K, Wetekam J (2024) FTIR spectroscopy and multivariate discriminant analysis for classifying bituminous mastics: Exploring aging states and mastic composition. Constr Build Mater 438. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2024.137188\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2024.137188\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMotevalizadeh SM, Mollenhauer K (2024) Exploration of chemical changes in bituminous mastics induced by aging: insights from FTIR spectroscopy, DSR measurements, and machine learning. Int J Pavement Eng. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.1080/10298436.2024.2418927\u003c/span\u003e\u003cspan address=\"10.1080/10298436.2024.2418927\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTauste R, Moreno-Navarro F, Sol-S\u0026aacute;nchez M, Rubio-G\u0026aacute;mez MC (2018) Understanding the bitumen ageing phenomenon: A review. Constr Build Mater 192:593\u0026ndash;609. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2018.10.169\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2018.10.169\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBehnood A, Modiri Gharehveran M (2019) Morphology, rheology, and physical properties of polymer-modified asphalt binders. Eur Polym J 112:766\u0026ndash;791. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.eurpolymj.2018.10.049\u003c/span\u003e\u003cspan address=\"10.1016/j.eurpolymj.2018.10.049\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBehnood A (2019) Application of rejuvenators to improve the rheological and mechanical properties of asphalt binders and mixtures: A review. J Clean Prod 231:171\u0026ndash;182. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jclepro.2019.05.209\u003c/span\u003e\u003cspan address=\"10.1016/j.jclepro.2019.05.209\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNam T, Taylor A, Willis R (2012) Effect of rejuvenator on performance properties of hma mixtures with high rap and ras contents\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin Zaumanis,Rajib RF (2013) B.Mallick, Evaluation of Rejuvenator\u0026rsquo;s Effectiveness with Conventional Mix Testing for 100% RAP Mixtures. Transp Res Board 1\u0026ndash;14\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNahar SN, Qiu J, Schmets AJM, Schlangen E, Shirazi M, Van De Ven MFC, Schitter G, Scarpas A (2014) Turning back time: Rheological and microstructural assessment of rejuvenated bitumen. Transp Res Rec 2444:52\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3141/2444-06\u003c/span\u003e\u003cspan address=\"10.3141/2444-06\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAli AW, Mehta YA, Nolan A, Purdy C, Bennert T (2016) Investigation of the impacts of aging and RAP percentages on effectiveness of asphalt binder rejuvenators. Constr Build Mater 110:211\u0026ndash;217. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2016.02.013\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2016.02.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbdelaziz A, Epps Martin A, Masad E, Ar\u0026aacute;mbula Mercado E, Kaseer F (2022) Effects of ageing and recycling agents on the multiscale properties of binders with high RAP contents. Int J Pavement Eng 23:1248\u0026ndash;1270. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10298436.2020.1797736\u003c/span\u003e\u003cspan address=\"10.1080/10298436.2020.1797736\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHolleran I, Masad E, Holleran G, Wubulikasimu Y, Malmstrom J, Wilson DJ (2021) Nanomechanical mapping of rejuvenated asphalt binders. Road Mater Pavement Des 22:2478\u0026ndash;2497. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14680629.2020.1771406\u003c/span\u003e\u003cspan address=\"10.1080/14680629.2020.1771406\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCavalli MC, Partl MN, Poulikakos LD (2019) Effect of ageing on the microstructure of reclaimed asphalt binder with bio-based rejuvenators, Road Mater. Pavement Des 20:1683\u0026ndash;1694. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14680629.2019.1594049\u003c/span\u003e\u003cspan address=\"10.1080/14680629.2019.1594049\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMenapace I, Garcia Cucalon L, Kaseer F, Ar\u0026aacute;mbula-Mercado E, Epps Martin A, Masad E, King G (2018) Effect of recycling agents in recycled asphalt binders observed with microstructural and rheological tests. Constr Build Mater 158:61\u0026ndash;74. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2017.10.017\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2017.10.017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu X, Zaumanis M, Dos Santos S, Poulikakos LD (2014) Rheological, microscopic, and chemical characterization of the rejuvenating effect on asphalt binders. Fuel 135:162\u0026ndash;171. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fuel.2014.06.038\u003c/span\u003e\u003cspan address=\"10.1016/j.fuel.2014.06.038\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbdelaziz A, Masad E, Epps Martin A, Mercado EA, Bajaj A (2021) Multiscale Characterization of Aging and Rejuvenation in Asphalt Binder Blends with High RAP Contents. J Mater Civ Eng 33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1061/(asce)mt.1943-5533.0003910\u003c/span\u003e\u003cspan address=\"10.1061/(asce)mt.1943-5533.0003910\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePipintakos G, Soenen H, Ching HYV, Vande Velde C, Van Doorslaer S, Lemi\u0026egrave;re F, Varveri A (2021) Van den bergh, Exploring the oxidative mechanisms of bitumen after laboratory short- and long-term ageing. Constr Build Mater 289. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2021.123182\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2021.123182\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMirwald J, Werkovits S, Camargo I, Maschauer D, Hofko B, Grothe H (2020) Understanding bitumen ageing by investigation of its polarity fractions. Constr Build Mater 250. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2020.118809\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2020.118809\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMastoras F, Varveri A, van Tooren M, Erkens S (2021) Effect of mineral fillers on ageing of bituminous mastics. Constr Build Mater 276. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2020.122215\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2020.122215\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBajaj A, Epps Martin A, King G, Glover C, Kaseer F (2020) Ar\u0026aacute;mbula-Mercado, Evaluation and classification of recycling agents for asphalt binders. Constr Build Mater 260. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2020.119864\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2020.119864\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa L, Varveri A, Jing R, Erkens S (2023) Chemical characterisation of bitumen type and ageing state based on FTIR spectroscopy and discriminant analysis integrated with variable selection methods. Road Mater Pavement Des. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14680629.2023.2181008\u003c/span\u003e\u003cspan address=\"10.1080/14680629.2023.2181008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeigel S, Stephan D (2018) Differentiation of bitumen according to the refinery and ageing state based on FTIR spectroscopy and multivariate analysis methods. Mater Struct Constr 51. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1617/s11527-018-1252-6\u003c/span\u003e\u003cspan address=\"10.1617/s11527-018-1252-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeigel S, Wetekam J, Mollenhauer K (2023) Identification and classification of PAH in asphalt binders with FTIR spectroscopy and multivariate analysis methods. Fuel 337. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fuel.2022.126845\u003c/span\u003e\u003cspan address=\"10.1016/j.fuel.2022.126845\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMotevalizadeh SM, Mollenhauer K (2024) Use of multivariate clustering analysis to investigate the physicochemical interactions in bitumen mastics using micromechanical modeling and FTIR spectroscopy. Constr Build Mater 448:138230. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.conbuildmat.2024.138230\u003c/span\u003e\u003cspan address=\"10.1016/j.conbuildmat.2024.138230\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGundla A, Underwood S (2017) Evaluation of in situ RAP binder interaction in asphalt mastics using micromechanical models. Int J Pavement Eng 18:798\u0026ndash;810. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10298436.2015.1066003\u003c/span\u003e\u003cspan address=\"10.1080/10298436.2015.1066003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWalther A, B\u0026uuml;chler S, Cannone Falchetto A, Wang D, Riccardi C, Wistuba MP (2019) Experimental investigation on asphalt mixtures prepared with reclaimed asphalt pavement and rejuvenators based on the BTSV method. Road Mater Pavement Des 20:1695\u0026ndash;1708. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14680629.2019.1594053\u003c/span\u003e\u003cspan address=\"10.1080/14680629.2019.1594053\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlisov A, Riccardi C, Schrader J, Cannone Falchetto A, Wistuba MP (2020) A novel method to characterise asphalt binder at high temperature. Road Mater Pavement Des 21:143\u0026ndash;155. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14680629.2018.1483258\u003c/span\u003e\u003cspan address=\"10.1080/14680629.2018.1483258\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchrader J, Wistuba MP, Falchetto AC, Riccardi C, Alisov A (2020) A new binder-fast-characterization-test using a dynamic shear rheometer and its application for rejuvenating reclaimed asphalt binder. J Test Eval 48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1520/JTE20180893\u003c/span\u003e\u003cspan address=\"10.1520/JTE20180893\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYaro AS, Maly F, Prazak P, Maly K (2024) Outlier Detection Performance of a Modified Z-Score Method in Time-Series RSS Observation With Hybrid Scale Estimators. IEEE Access 12:12785\u0026ndash;12796. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ACCESS.2024.3356731\u003c/span\u003e\u003cspan address=\"10.1109/ACCESS.2024.3356731\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith B (1998) Infrared spectral interpretation: a systematic approach. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.1201/9780203750841\u003c/span\u003e\u003cspan address=\"10.1201/9780203750841\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBakhtiari MT (2015) Role of Sodium Hydroxide in Bitumen Extraction: Production of Natural Surfactants and Slime Coating, 206\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXiaolong Zou XW, Xie H, Li J, Jing H, Li H, Li Z (2024) Innovative application of coffee grounds oil as an asphalt modifier: Extraction, preparation, and rheological properties, Case Stud. Constr Mater 21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.1016/j.cscm.2024.e04012\u003c/span\u003e\u003cspan address=\"10.1016/j.cscm.2024.e04012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoise V, Caputo P, Porto M, Calandra P, Angelico R, Rossi CO (2019) A review on Bitumen Rejuvenation: Mechanisms, materials, methods and perspectives. Appl Sci 9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/app9204316\u003c/span\u003e\u003cspan address=\"10.3390/app9204316\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHesse M, Meier H, Zeeh B (2005) \u0026ndash;7 Spektroskopische Methoden in der organischen Chemie Auflage), 7th ed., Georg Thieme, Stuttgart, 2005\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRen S, Liu X, Lin P, Jing R, Erkens S (2022) Toward the long-term aging influence and novel reaction kinetics models of bitumen. Int J Pavement Eng 24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10298436.2021.2024188\u003c/span\u003e\u003cspan address=\"10.1080/10298436.2021.2024188\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu X, Colman SM, Brown ET, Minor EC, Li H (2013) Estimation of carbonate, total organic carbon, and biogenic silica content by FTIR and XRF techniques in lacustrine sediments. J Paleolimnol 50:387\u0026ndash;398. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10933-013-9733-7\u003c/span\u003e\u003cspan address=\"10.1007/s10933-013-9733-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMirwald J, Nura D, Hofko B (2022) Recommendations for handling bitumen prior to FTIR spectroscopy. Mater Struct Constr 55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1617/s11527-022-01884-1\u003c/span\u003e\u003cspan address=\"10.1617/s11527-022-01884-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHofer K, Mirwald J, Maschauer D, Grothe H, Hofko B (2022) Influence of selected reactive oxygen species on the long-term aging of bitumen. Mater Struct Constr 55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1617/s11527-022-01981-1\u003c/span\u003e\u003cspan address=\"10.1617/s11527-022-01981-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University of Kassel","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"FTIR Spectroscopy, Multivariate Discriminant Analysis, Machine Learning, Classification, Partial Least-Square Regression, Linear Discriminant Analysis","lastPublishedDoi":"10.21203/rs.3.rs-6342505/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6342505/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study examines the rejuvenation mechanisms in bituminous RAP mastics using Fourier transform infrared (FTIR) spectroscopy and multivariate discriminant analysis. The chemical interactions between RAP binders, virgin bitumen, warm mix additives, and two commercial recycling agents (RJs) were evaluated to determine their effectiveness in restoring aged binders. A hybrid approach combining partial least squares regression and linear discriminant analysis (PLSR-LDA) was applied to extract latent variables, classify samples, and identify critical wavenumbers associated with rejuvenation. The findings indicate that Sylvaroad mitigates oxidation effects, particularly around ~\u0026thinsp;1758\u0026ndash;1715 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, while Storflux primarily influences methyl bending near 1370\u0026ndash;1360 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e. Despite these effects, irreversible oxidative aging remains evident in the 1300\u0026ndash;1160 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e range, linked to oxygen-containing compounds, suggesting that complete restoration of binder properties is unattainable. Additionally, mineral composition markers, particularly in the 489\u0026ndash;446 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and 590\u0026ndash;555 cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e ranges, persist as key indicators for distinguishing varying RAP contents after rejuvenation.\u003c/p\u003e","manuscriptTitle":"Rejuvenation mechanisms in bituminous RAP mastics: insights from FTIR spectroscopy and multivariate discriminant analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-01 09:41:08","doi":"10.21203/rs.3.rs-6342505/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a58ab6b5-9aeb-41ab-9050-135cfbb0aec8","owner":[],"postedDate":"April 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":46436156,"name":"Civil Engineering"}],"tags":[],"updatedAt":"2025-04-01T09:41:09+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-01 09:41:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6342505","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6342505","identity":"rs-6342505","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}