During the mixing of rubber with formulation components or after the vulcanization process, whitish spots often appear on the surface of the rubber (Arabit & Pajarito, 2015; Huang et al., 2019; Ignatz-Hoover & Datta et al., 2003; Choi, 1999; Choi, 2004; Choi et al., 2009; Nah & Thomas, 1980). This phenomenon, commonly referred to as additive blooming or vulcanization component blooming, results from the migration of certain formulation components to the rubber surface. In general, the literature reports that rubber and some vulcanization components exhibit partial incompatibility (Ciesielski, 1999).

Thus, the blooming phenomenon on the surface of rubber can occur when an additive from the formulation dissolves in the rubber matrix under the elevated temperatures of the mixing and vulcanization processes. After these processes, part of the additive tends to migrate out of the polymer matrix and crystallize on the rubber surface (Guo et al., 2008; Saeed et al., 2011; Munasinghe, 2004). The resulting crystals consist of solidified additives that appear as a whitish or powder-like deposit accumulated on the surface.

This phenomenon adversely affects the aesthetic appearance and overall quality of the final product. Specifically, certain formulation components tend to migrate out of the polymer matrix due to their incompatibility with the rubber, and this effect becomes more pronounced over time. The blooming behavior of these components has been widely studied to identify the substances that migrate to the surface and to develop strategies for minimizing this effect. Techniques such as Fourier Transform Infrared Spectroscopy (FTIR) and Attenuated Total Reflectance Infrared Spectroscopy (ATR-IR) have been employed for this purpose (Santiago et al., 2016; Torregrosa-Coque et al., 2011). Although several studies have investigated the surface blooming of individual formulation components in vulcanized rubbers, few have examined the dominance or interaction among different blooming species.

The literature reports the use of fillers with inhibitory properties, such as graphene and silica, to mitigate the surface blooming phenomenon in polymeric composites (Huang et al., 2019; Saeed et al., 2012; Xing et al., 2017; Garing et al., 2023). However, some of these fillers have been found to adversely affect the internal structure of natural rubber composites. Blooming is a critical phenomenon that influences the overall properties and performance of rubber materials. Surface blooming occurs most frequently in both natural and synthetic vulcanized rubber compounds. Although the problem is primarily aesthetic, it is also believed that these migrated additives may alter the mechanical properties of the final product. Previous studies have shown that the concentration of additive components in the rubber formulation significantly affects both the extent and rate of blooming (Pajarito et al., 2014; Arabit & Pajarito, 2019).

(Wang et al., 2024) investigated the formation of surface precipitates and the changes in the tensile properties of nitrile butadiene rubber (NBR) after storage at 85 °C in air. They concluded that the reaction between zinc oxide and stearic acid under prolonged thermal exposure promotes blooming and contributes to the degradation of NBR’s tensile properties. A pronounced decrease in the elongation at break (δk) was observed, which was correlated with the blooming of zinc stearate and its consequent reduction of the plasticizing effect.

(Anozie et al., 2025) studied the effect of alginate bead–encapsulated sulfur on the blooming characteristics of natural rubber (NR), comparing their behavior with that of soluble and insoluble sulfur. At test temperatures of 100 and 120 °C, they concluded that sulfur encapsulated in alginate beads produced significantly less blooming than soluble sulfur. Statistical analysis showed that the tensile strength of vulcanized rubber containing microspheres (18.9 ± 2.5 MPa) did not differ significantly (p = 0.067) from that of rubber with insoluble sulfur (22.2 ± 2.5 MPa), although it was lower (p = 0.011) than that of rubber with soluble sulfur (23.3 ± 0.7 MPa). The authors concluded that the use of encapsulated sulfur effectively controls blooming without significantly compromising the mechanical properties compared to insoluble sulfur.

Krieg et al. (2021) presented a visual catalog of damage caused by degradation phenomena in rubber materials. The analysis of several cases of blooming enabled the identification of different compositions and origins, demonstrating that describing a degradation phenomenon is only the first step toward understanding it. Using Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR-FTIR) as a surface characterization technique, the authors emphasized the need for further research on damage phenomena to better elucidate their causes and the structural changes occurring within the material. Such understanding would facilitate the design of preventive conservation strategies and interventions, ultimately contributing to a more sustainable relationship with these materials at both industrial and societal levels.

Basso et al. (2021) investigated the feasibility of using ground tire rubber (GTR) as a raw material for injection molding by blending it with polypropylene (PP). Nine different formulations were prepared by varying the GTR-to-PP ratio and employing three types of rubber powder. The authors concluded that high contents of rubber powder-up to 80% recycled material can be successfully incorporated into molding processes. This approach reduces the need for virgin polymers while increasing the proportion of recycled material, thereby contributing to more sustainable manufacturing practices.

The incorporation of recycled rubber as a formulation component in vulcanized rubber products has been increasingly explored in recent years (Ren et al., 2023; Ma et al., 2021; de Souza et al., 2019; Song et al., 2018). The addition of reclaimed rubber into new formulations represents a significant advancement toward sustainable manufacturing in the rubber industry. This approach not only adds value to waste materials but also promotes environmentally responsible recycling practices, supporting the development of high-performance products with a reduced environmental footprint.

Although several methods for inhibiting blooming have been investigated, there is still a gap in quantitative studies that systematically correlate the intensity of blooming with the evolution of mechanical property degradation in compounds containing high amounts of ground tire rubber (GTR).

On the other hand, the Response Surface Methodology (RSM) comprises a set of mathematical and statistical techniques designed to model and analyze engineering processes in which a response of interest is influenced by multiple independent variables. The main objective of RSM is to optimize this response, which typically represents a critical quality attribute of a product or process (Raissi & Farsani, 2009). In many manufacturing systems, performance is characterized by several interrelated response variables. Optimizing such systems therefore requires complex decision-making to simultaneously improve all relevant responses a challenge formally referred to as a multi-response optimization problem (Cabrera-Castro, 2014).

In RSM, the relationship between the independent variables (x1, x2, ..., xn) and the response “y” is represented as in Eq. (1):

where

For multi-response optimization, a common and effective strategy involves the use of a desirability function, which enables the simultaneous optimization of several responses by converting them into a single composite metric. This approach has been widely adopted across diverse engineering and materials science applications. For instance, (Okuma et al., 2022) applied RSM to optimize welding parameters current, voltage, and travel speed and reported a strong correlation between predicted and experimental ultimate tensile strength (UTS) values, with optimal results ranging from 425 to 475 MPa. Similarly, (Boyacı and Baynal, 2019) employed RSM with a Box–Behnken design to optimize vulcanization parameters (time, temperature, and pressure) in elastomeric bearings, identifying optimal conditions of 146.17 min, 112.92 °C, and 172.72 bar to achieve target compressive stiffness and deflection values.

Further applications of the response surface methodology can be found in the work of (Aktar et al., 2024), who optimized a chloroprene/natural rubber (CR/NR) blend for automotive applications. Their study identified that a natural rubber proportion of approximately 27%, combined with specific additive levels, optimized key performance properties such as hardness and tensile strength. Similarly, (Paredes et al., 2021) optimized the mechanical properties of a vulcanized rubber material, determining an optimal formulation and curing condition 6 minutes at 165 °C that maximized overall desirability.

The effectiveness of RSM coupled with the desirability function has been consistently demonstrated across various engineering domains. In the materials sector, (Thongchom et al., 2023) applied this approach to optimize the mechanical behavior of polypropylene-based nanocomposites reinforced with graphene, basalt, and ethylene propylene diene monomer (EPDM). They found that incorporating approximately 1 wt% graphene and 25 wt% basalt yielded optimal mechanical performance, enhancing tensile strength, modulus of elasticity, and impact resistance by 21%, 58%, and 15%, respectively.

Another innovative application of RSM was reported by Anjali and Venkatesan (2024), who optimized a concrete composite incorporating recycled tire rubber (crumb rubber) for construction purposes. By combining response surface methodology with a hybrid neural network model, they identified that the formulation containing aluminum, hydrogen peroxide, and 2.5% rubber (Al+H₂O₂+2.5R) achieved superior optimization of mechanical performance, attaining compressive and tensile strengths of 46.08 MPa and 3.80 MPa, respectively.

From a process optimization perspective, Ghumman et al. (2021) applied RSM to optimize the reverse vulcanization reaction between sulfur and rubber seed oil, aiming to minimize unreacted sulfur content in the final copolymer. Their findings revealed that the initial sulfur content was the most influential factor. Although optimization reduced residual sulfur to 5.8%, complementary strategies such as tetrahydrofuran filtration and the incorporation of 5–10% crosslinkers, further decreased it to 0.25%. The resulting terpolymers exhibited enhanced thermal stability and an amorphous morphology, as confirmed by FESEM and p-XRD analyses.

This study presents a comprehensive strategy for the multivariable optimization of recycled rubber compounds, employing Response Surface Methodology (RSM) combined with desirability functions to simultaneously enhance mechanical performance and minimize surface blooming. A key innovation of this work is the development of a novel technique for quantifying blooming using a vision system coupled with image processing, which measures bloom by calculating the Euclidean distance between RGB color layers over time. This method effectively distinguishes progressive additive blooming from initial surface defects, enabling precise monitoring throughout the process. The proposed system represents a significant advancement for quality control in the vulcanized rubber industry and provides a robust digital tool that can be readily integrated into Industry 4.0 frameworks for automated process optimization and control.

Figure 1 illustrates the methodological framework adopted in this study, encompassing all stages of the experimental procedure from the preparation of the composite materials to the final optimization process. The workflow begins with the mixing of formulation components, followed by the vulcanization of each sample. Subsequently, a vision system was developed to estimate the degree of blooming on the rubber surface, after which the mechanical properties of each specimen were measured to construct a comprehensive dataset. A total of 72 experimental runs were conducted according to a central composite design, ensuring statistical robustness and reproducibility. This dataset was then employed to perform the statistical analysis and multivariable optimization of the key process parameters.

The design and analysis of experiments play a crucial role in engineering research, as they enable systematic improvement of process and system performance (Cabrera-Castro, 2014). From an engineering standpoint, enhancing the durability and fatigue resistance of rubber materials is essential, given that elastomeric components are often exposed to extreme conditions of frequency, deformation, and temperature (Schneider et al., 2023). Therefore, developing GTR-based rubber composites with superior mechanical strength and stability is of significant practical relevance.

A multi-criteria optimization approach was implemented to evaluate the individual and interactive effects of formulation factors on the blooming response. To reconcile the potentially conflicting influences among these factors and determine a balanced operating condition, the desirability function methodology was applied. This method converts individual responses each expressed in different physical units into a dimensionless desirability scale, facilitating simultaneous optimization. The approach enables the identification of optimal operating levels for all evaluated factors, thereby minimizing the occurrence of surface blooming in vulcanized rubber systems. In this methodology, the predicted response values y (yij^), are transformed into corresponding desirability values (d_ij), according to the range (0 ≤ dij ≤ 1), These individual desirabilities are then combined using a geometric mean to obtain an overall desirability index. The value of “d” represents the proximity of the predicted response to its target, where 1 corresponds to the ideal condition and 0 indicates complete undesirability (Raissi & Farsani, 2009).

The simultaneous optimization of multiple responses was performed using this desirability framework, following the criteria summarized in Table 4. Specific limits, targets, and weighting factors were defined for each response variable: the blooming phenomenon was set to be minimized, while elongation at break and tensile strength were maximized. Conversely, Young’s modulus was optimized to reach a specific target value. Two weighting parameters were defined to control the relative importance and the curvature of each individual desirability function. A uniform impact factor of 3 was assigned to all responses to reflect their comparable relevance in the overall optimization process.

This optimization strategy effectively prioritized efflorescence reduction during the multivariable optimization process, which consequently led to improvements in product yield, mechanical balance, and surface quality. The analysis identified a global optimum within the design space, corresponding to run 12, with a predicted overall desirability of 0.753. The specific combination of factor levels and operating conditions associated with this optimum is summarized in Table 5a.

Upon experimental validation, the observed desirability reached 0.829, exceeding the predicted value and thereby confirming the robustness and reliability of the proposed regression models. The predicted optimum responses for blooming, elongation, Young’s modulus, and tensile strength are shown in Table 5b, demonstrating close agreement between experimental and predicted values, which validates the model’s accuracy in capturing the complex nonlinear relationships among formulation components.

From a formulation standpoint, the optimal conditions correspond to a balanced combination of accelerator and vulcanizing agent levels that favor a homogeneous crosslink distribution. This balance effectively limits additive migration toward the surface (reducing blooming) while maintaining adequate chain mobility for improved flexibility and tensile performance. The results indicate that moderate accelerant levels, coupled with controlled sulfur content, promote efficient vulcanization kinetics without over-curing, thereby achieving a favorable trade-off between stiffness and elasticity.

In practical terms, this optimization provides a scientifically grounded formulation strategy for GTR-based vulcanizates, capable of minimizing surface blooming without compromising mechanical integrity. The results highlight the potential of response surface methodology combined with desirability analysis as a powerful tool for tuning multifunctional rubber systems, particularly in applications where both aesthetic surface quality and mechanical durability are critical

The optimal formulation predicted by the model established the following coded levels for the variables: additive (CCA1) = –0.345 phr, activator (CCB1) = –0.979 phr, accelerator [CCD2] = –0.632 phr, and vulcanizer (S) = 0.479 phr. The fixed formulation parameters were maintained at (CCA2) = 10 phr, [CCB2] = 1.25 phr, (CCD1) = 1.35 phr, (CCC1) = 1.75 phr, and (GTR) = 100 phr. Notably, the negative coded values of (CCA1), (CCB1), and (CCD2) correspond to experimental conditions where no additional quantities of these components were required. This finding suggests that the intrinsic composition of the recycled rubber (GTR) already provides a sufficient baseline of these additives for adequate processing, thereby eliminating the need for supplementary addition. The identified combination of factors effectively minimizes surface blooming while maintaining the desired mechanical balance in the vulcanized compound, demonstrating the efficiency of the optimized formulation.

Figure 7 illustrates the multivariable optimization outcome derived from the global desirability function (Figure 7a), which simultaneously integrated the responses of blooming, tensile strength, elongation, and Young’s modulus. The contour graph represents the variation of these properties as a function of the additive (CCA1) and vulcanizer (S), while maintaining (CCB1) and (CCD2) constant at –0.891 and –0.657, respectively.

The overlap analysis identified a feasible optimization region (highlighted in yellow) where all performance criteria are simultaneously satisfied. Within this region, the global optimum was located at the coded coordinates X₁ = –0.331 (CCA1) and X₂ = 0.676 (S), predicting the following responses: blooming = 0.0153 (minimized), tensile strength = 7.01 MPa, elongation = 92.9%, and Young’s modulus = 8.14 MPa. This combination of factors represents the best compromise between surface stability and mechanical reinforcement, yielding a high overall desirability value.

As shown in Figure 7b, the blooming phenomenon remains at a low-to-moderate level within this optimal region, avoiding the peaks associated with excessive additive or vulcanizer concentrations. These findings confirm that the proposed optimization strategy successfully identified robust processing conditions that minimize surface migration while preserving the mechanical integrity and functional performance of the recycled rubber composite.

To verify the reliability of the predicted optimal formulation, experimental validation was carried out under the factor levels identified by the global desirability analysis. The experimental results were compared with the model predictions and the reference conditions (prior to optimization) to assess the consistency and predictive accuracy of the optimization strategy. Figure 8(a) shows two groups of blooming profiles of the system before optimization, where the untreated formulations exhibited high variability and marked surface blooming, indicating low compatibility of the components and non-uniform cross-linking. In contrast, Figure 8(b) shows the validated optimized formulation, which demonstrates a substantial reduction in blooming intensity and, consequently, greater mechanical uniformity. The visual and quantitative comparison between the two figures clearly confirms the effectiveness of the optimization model in identifying the processing conditions that minimize blooming and improve mechanical stability.

The experimental results closely matched the model predictions, with deviations of less than 5% for all response variables. Specifically, experimental desirability reached 0.829, compared to the predicted value of 0.753, confirming the robustness and predictive reliability of the developed model.

Overall, these results demonstrate that the combined use of response surface methodology (RSM) and global desirability optimization provides a robust and experimentally validated framework for the precise adjustment of vulcanized rubber compound formulations containing ground tire rubber (GTR). This strategy not only improves product quality and reproducibility, but also reduces additive consumption and process variability, offering a scalable approach to sustainable rubber recycling and formulation control.

This study successfully demonstrates the development and application of a comprehensive methodology for analyzing, quantifying, and predicting the blooming phenomenon in vulcanized rubber composites containing ground tire rubber (GTR). The main contribution lies in the implementation of a non-destructive measurement system based on digital image processing, which, by calculating the Euclidean distance between RGB color spaces over time, allows for the reliable and objective quantification of the progressive surface migration of additives, discriminating between the initial texture and subsequent aesthetic degradation. The proposed method distinguishes between initial surface characteristics and additive blooming, providing a reliable tool for monitoring the aesthetic degradation of vulcanized rubber.

The system developed not only facilitates quantification, but also facilitates the modeling of blooming using regression variables, enabling prediction and control in future formulations. Using a multivariable optimization approach based on the Response Surface Methodology (RSM), an optimal formulation was identified that significantly minimizes blooming without compromising key mechanical properties (tensile strength, elongation, and Young’s modulus). Contrary to expectations, the optimal condition corresponds to the absence of supplementary addition of the additives studied, indicating that the residual load present in the GTR is sufficient. This crucial finding suggests that excessive dosing is not only unnecessary but is the main contributing factor to surface blooming.

From a formulation standpoint, the optimal conditions correspond to a balanced combination of accelerant and vulcanizing agent levels that favor a homogeneous crosslink distribution. This balance effectively limits additive migration toward the surface (reducing blooming) while maintaining adequate chain mobility for improved flexibility and tensile performance. The results indicate that moderate accelerator levels, coupled with controlled sulfur content, promote efficient vulcanization kinetics without over-curing, thereby achieving a favorable trade-off between stiffness and elasticity.

The optimized formulation, experimentally validated with excellent agreement between predicted and observed values (global desirability = 0.829), revealed that reducing additive and accelerator concentrations yields substantial benefits. This approach not only limits additive migration and blooming intensity but also maintains desirable levels of tensile strength, elongation, and Young’s modulus. The validation experiments confirmed that the proposed model exhibits high predictive accuracy (error < 5%), supporting its robustness and applicability to industrial-scale processes. Economically, the optimized formulation leads to significant cost savings (≈15–20%) through reduced additive consumption and lower defect rates. Industrially, the integration of the vision-based blooming quantification system provides a non-destructive, automated, and scalable quality control tool, facilitating real-time process monitoring within Industry 4.0 frameworks. The combination of statistical design, machine vision, and multivariable optimization thus represents a comprehensive digital manufacturing strategy that enhances both process efficiency and material sustainability. Despite the robustness and predictive accuracy of the optimization model, several limitations should be acknowledged. First, the experimental scope was limited to four formulation factors (additive, activator, accelerant, and vulcanizing agent) under controlled laboratory conditions. While this selection captured the main effects governing blooming and mechanical properties, other influential variables such as mixing temperature, shear rate, curing pressure, and particle size distribution of GTR were explicitly included. But is possible may be that other components of the formulation affect the migration kinetics of additives and crosslink network formation.

Second, the vision-based quantification system developed in this study which, although effective for surface blooming detection, does not provide chemical composition data about the migrated species. Complementary characterization techniques such as FTIR spectroscopy, XPS, or ATR-IR could enhance the understanding of the underlying physicochemical mechanisms of efflorescence and its relationship to additive polarity and diffusion.

Finally, future research should evaluate the long-term aging behavior and recyclability of the optimized compounds to ensure sustained mechanical integrity and surface stability throughout the product’s service life, thus strengthening their potential for large-scale implementation within the circular economy of rubber materials. Overall, this research establishes a quantitative and reproducible methodology for balancing performance, cost, and environmental impact in recycled rubber formulations. The proposed approach serves as a reference model for future developments in circular polymer technologies, advancing the sustainable reuse of end-of-life tire materials through science-based process optimization.