Material processing conditions effect on the thermokinetics of end of life tyres

ABSTRACT In this work, end of life tyres (ELTs) samples were subjected to pyrolysis in a dynamic thermogravimetry set-up under various heating rates between 5°C and 25°C min−1. The experimental design involved dried and non-dried sample results in an isothermal step within the heating program, to be able to compare the data against non-dried samples, to further investigate the effect of drying on the results obtained. The drying of the samples resulted in an increase in the onset temperature by some 30°C. However, cryomilling the ELTs samples showed a great effect in reducing the onset of the degradation curve by about 10°C amongst the same set of experiments. The type of sample, in terms of size and processing/milling condition, had a significant impact on the heat flow properties and calorimetry reported. The exothermic region for the original shavings resulted in an enthalpy of −48 J g−1, whilst the cutting mill samples exhibited an enthalpy increase to −184 J g−1. However, the cryomilled samples showed an enthalpy increase estimated at 139 J g−1.


Introduction
The increase in global population has resulted in a high demand on access to high-quality roads and transport. Hence, an increase in the production of tyres has been noted as of late which coincides with hefty proportions of rubber waste and end of life tyres (ELTs) components in the solid and special waste streams on a global scale (Formela 2021). About 1.6 billion tyre units are produced annually but only 100 million units are processed by the recycling industry around the world (Jansen, van der Walt, and Crouse 2022). This can only lead to more burdens on the surrounding urban environment when ELTs are not disposed of correctly, considering that improper disposal, stockpiling and landfilling of such type of waste can result in rodent infestation, pollution of aquifers and aquatic environments, as well as, leaching of toxic chemicals (Ferdous et al. 2021). There exists a number of ELTs management options that are applied on varies scales in various regions across the globe. These include recycling/ physical and mechanical treatment, use in civil engineering purposes, use as a tyre derived fuel in incineration units and recovery of valuable chemicals via pyrolysis (Campbell-Johnston et al. 2020). Major criticism is faced by advocates of ELTs incineration whom argue that releasing the high energy content of ELTs (>30 MJ kg −1 ) can reduce reliance on classical fossil fuels (USEPA 2016). However, incineration of ELTs is also associated with high rates of dioxins, sulfur oxides, polycyclic aromatic hydrocarbons, fine particulates emissions (Arabiourrutia et al. 2020). Pyrolysis on the other hand produces valuable products comparable to fuels and can easily be marketed with high environmental off-set potential and associated incidences (Al-Salem, Lettieri, and Baeyens 2009). Pyrolysis is defined as a thermo-chemical conversion (TCC) process that takes place in inert atmospheres and cracks tyres using operating temperatures reaching to 800°C; which results on the other hand in the production of high-value oils, light/permanent gases that could be used for heating purposes and solid char/carbon black (Dwivedi, Manjare, and Rajan 2020). A large bulk of technical literature is published and supports the process of pyrolysis (Hoang, Nguyen, and Nguyen 2020), and has emphasized that experimental set-up (Osorio-Vargas et al., 2012) and type of reactor are key elements for producing high end value products, (Januszewicz et al. 2020). The reader is also referred to Zhang et al. (2022) for the classification of pyrolysis technologies and types.
A major emphasis in research and development within the past decade or so is dedicated to upgrading ELTs pyrolysis products namely derived oils (i.e., pyro-oils). Techniques on pre-and posttreatment have been considered for such which include de-sulphurisation, drying and moisture removal and hydro-treatment (Aydin and Ilkilic 2012). However, drying and pre-conditioning (e.g., milling and size reduction) as a pre-treatment step are a scant topic in literature and data resulting from experiments of such are quite scarce (Danon et al. 2015a). In addition to past studies associated with properties of the pyrolysis process products, their exist a large volume of literature that report the kinetics of the degradation reaction of ELTs in various scales without linking it to pre-treatment conditioning or size reduction. Cheung et al. (2011) showed that pyrolysis of ELTs is an endothermic reaction linked directly to the operating temperature and heating rate. The energy usage model previously developed by Lam, Lee, and Hui (2010) showed that multi-stage pyrolysis approach can reduce energy requirements by up to 21.7%. It is essential to study the thermal and calorimetric properties and associated kinetics (thermokinetics) of various sizes of feedstock treated and report on the apparent activation energy (E a ) associated with such reactions. This is considered the cornerstone in designing the appropriate units and determine whether such conditioning and treatment steps have an impact on the quality of the products of pyrolysis (Arabiourrutia et al. 2019). Previous attempts to determine the thermokinetics of various materials was reported in literature determining the kinetics triplets using a number of analytical methods. Fernandez et al. (2022) studied the pyrolysis and gasification kinetics of the residues of agro bio-waste. Coats and Redfern method was used to determine the mechanism of degradation resulting in a first-order behavior. Torres-Sciancalepore et al. (2022) studied the thermokinetics and calorimetric values of both quince and pectin-free quince waste, determining that the latter had a higher activation energy with a better regression fit using the Flynn-Wall-Ozawa method.
In this work, ELTs are subjected to thermogravimetry and differential scanning calorimetry to determine the properties of three batches of rubber waste, namely shavings from tyre carcasses, powder resulting from a cutting mill process and another from cryogenic milling (i.e., cryomill). The E a values was estimated using various methods to determine the effect of particle size change. It is essential to understand the thermal properties to be able to determine future directions of scale-up options and investments associated with such TCC units and integration options with existing industries. To best of knowledge, no previous attempt was made to detail the results of thermokinetics resulting from different pre-treatment and conditioning of ELTs feedstock. The work presented herein also shows in detail the impact of different conditioning and treatment on the resulting products of pyrolysis. This can be used in the future for unit and operation design of units dedicated to such processes, in an effort to reduce environmental burdens of ELTs waste accumulation in urban environments.

Samples
Scrap cars ELTs were acquired from a local dealership (165/65R14-State of Kuwait) which the sidewalls and carcass were exposed to the outdoors for a period of < two weeks after reclamation. The tyres bodies were used in this work after air blowing and shredding using an ELDAN Co. (Denmark) at 60•C to obtain rubber shavings from the rubber carcass of an approximate size of 1.15 cm (steel free).
Further details could be found in Al-Salem (2022) and the physcio-chemical properties and elemental analysis (EA) are depicted elsewhere including analytical procedures used in Al-Salem (2021a;2021b;2022) which are summarized in Tables S1-S2 of the Supplementary Materials File. Figure S1 also shows a pictorial depiction of the material in its original shavings form used in this work.

Milling and size reduction
In order to study the ELTs with respect to their size and as per the size reduction method, the samples were separated into three batches. The first was kept as is and will be referred to as original shavings. The second batch was subjected to milling using a cutting mill machine and the third to cryogenic milling (cryomill) as detailed herein. This will enable the investigation on the effect of the history of each type on the thermal and chemical properties investigated at later stage. A Fritsch PULVERISETTE 19 cutting mill was used with a 50 Hz frequency with an RPM between 1,168 and 2,800 to mill the tyre samples using a 0.5 mm mesh for a duration of 5 minutes in a single cycle (see Figure S2). Overheating was avoided to eliminate any evolution of gaseous from the decomposition of the rubber fraction from the tyre shavings. The materials were also subjected to cryomilling using a Retsch cryomill equipped with a 50 L LN tank (stainless steel grinding ball size = 25 mm, 50 ml grinding jar) using the following program: Number of cycles (6) in two consecutive runs, Pre-cooling time and frequency (10 min and 10 Hz), Cycle time -grinding (2 minutes and 25 Hz) and intermediate cooling (1 min and 5 Hz). Table S3 depicts the conditions of the mill used which were unified to reflect the best quality control practice in the laboratory. Figure S3 shows the cryogenic milling assembly used and a sample of the milled materials. The materials produced were of an average size of 85 × 46 mm of the cutting milled samples and 0.66 × 0.33 mm for the cryomilled samples verified by placing the materials under A TESCAN-VEGA 3 model scanning electron microscope equipped with a Bruker Energy Dispersive Elemental Spectroscopy (EDS) analysis system and software, which was utilized in this study ( Figure S4). Imaging was conducted using a voltage around 30 kV (to prevent damage of specimen) resulting in various micrographs taken with a 100 to 10,000-fold magnification. The sample surfaces were placed on a double-sided carbon tape positioned on a metal holder and then coated with a layer of gold with a thickness of 5 nm by sputtering using a JEOL-JFC-1600 Auto Fine sputter coater for 30 s using Argon gas with a flowrate of 5 nm min −1 and a pressure 3 × 10 −1 mbar (see Figures S5-S6 and Table S4).

Thermogravimetric Analysis (TGA) on rubber samples
The thermal behavior and subsequently the deactivation energy and associated calculations were investigated using thermogravimetric analysis (TGA) in this work. The TGA tests were run on all samples in duplicate at following heating rates: 5°C, 10°C, 15°C, 20°C, and 25°C min −1 on Netzsch TG 209 Tarsus system and analyzed using Proteus 8.0 software. Specimens were tested as supplied and after drying to investigate further the conditioning effect on the samples (Danon et al. 2015a) using the test programs defined below in Table 1.
Specimen size used was approximately 5 mg. For samples in form of shavings, specimens were taken from rubber pieces avoiding the shredded fibers. The test pan used was an aluminum oxide (open pan) for a sample weight of 5 mg in duplicates. The TGA unit is externally calibrated and showing the weight loss of sample against the temperature. Using the Netzsch Proteus 8.0 analysis software, the following properties were measured from the thermogravimetry (TG) thermograms (Yusriah et al. 2014): i) Onset decomposition temperature: measured as the temperature of 5% weight loss by onset function of the Proteus software and was defined as T os , ii) Temperature at 50% weight loss: the temperature at 50% weight loss was measured using cursor selection of the software (T 50% ), iii) Max decomposition temperature: the temperature at maximum weight loss was measured using cursor selection (T max ), iv) Inflection point temperature: measured using inflection function of the Proteus software (T if ), v) Residual mass %: measured using residual mass % function of the Proteus software; and, vi) The TGA curves from two specimens at a set heating rate and condition (dried or non-dried) were averaged and presented in the nest section.

Thermogravimetric analysis coupled with Gas Chromatography-Mass Spectroscopy (TG-GCMS) on rubber samples
Specimen preparation consisted of cutting a piece of about 10 mg from the rubber samples (avoiding fibers and other particles) and placing inside an alumina crucible (with another empty alumina crucible used as a reference pan). Simultaneous thermogravimetry and differential scanning calorimetry (TGA/DSC) was conducted under argon atmosphere flow rate of 30 ml min −1 using Perkin Elmer STA 6000. Each sample was heated from 30°C to 110°C at a scan rate of 10°C min −1 followed by a drying period of 5 min at 110°C. The temperature was then ramped up to 600°C with a heating rate of 10°C min −1 (Mkhize et al. 2019). Gases formed during heating were then transferred through a transfer line set at 250°C coupled to a GC-MS system (Clarus 680 GC and Clarus SQ MS) at a flow rate of 100 ml min −1 . The GC temperature gradient was ramped from 80°C to 280°C at a rate of 10°C min −1 . The temperature was then held at 280°C for 10 min. The total duration of the GC analysis was 30 min. The mass spectrometer was operated in the positive electron ionization (EI) mode, the electron energy was 70 eV, and the filament current was 270 mA. The mass range was 50 to 300 m/z without solvent delays. NIST mass spectral search program V.2.2. was used to assign detected GCMS peaks.

Degressive kinetics investigation
A detailed degressive reaction kinetics analysis was conducted on the three batches of ELTs under investigation (e.g., original shavings, cutting mill and cryogenically milled samples) with respect to the dataset extracted as well for dried and non-dried testing protocols. This was conducted to determine the influence of drying and particle size due to various processing conditions on the estimated apparent activation energy in this work (E a ). The analysis conducted was in-line with the International Confederation for Thermal Analysis and Calorimetry (ICTAC) recommendations and protocols (Vyazovkin et al. 2014). We first define the mass fraction as (α) thus to establish the derivation of each kinetic method under consideration.
where m, m o, and m f are the mass of the material at a specific reaction time (t), and initial and final mass of the rubber sample at the investigated temperature, respectively (Al-Salem and Khan 2014; Al-Salem et al. 2018). Rate of degradation in the irreversible thermal cracking reaction was defined with respect to the rate of reaction by incorporating the first-order Arrhenius equation as per the following (Vyazovkin et al. 2011): where A o stands for the Arrhenius factor (min −1 ), E a is the apparent activation energy (kJ mol −1 ), T is the reaction temperature at desired time (K), and f(α)is the solid-state reaction model with respect to the reaction order (n) (Al-Salem et al. 2017). The first kinetics evaluation method used was the Friedman's isoconversional method . Isoconversional methods have also been recently applied to determine the order and mechanism of degradation reaction (Saffe et al. 2019). It should be strictly noted that isoconversional methods are commonly applied, in order to, evaluate the degressive kinetics and are best applied when the kinetic triplet of the reaction is easily determined (Vyazovkin et al. 2011). The natural logarithm is applied to Eq.. (2) with respect to a fixed heating rate (β) and conversion intervals of 5% for a more accurate estimation. The derived equation will be as per the following for each material type under investigation.
Thus, the Friedman's plot will reflect ln dα = dt � � versus 1/T at a fixed value of conversion which its slope will equal to of E a /R and E a is estimated at each β (Al-Salem et al. 2018). The Flynn-Wall-Ozawa (FWO) isoconversional method was also applied to estimate the E a values for the degradation reactions in accordance with the methodology described previously by Poletto, Zattera, and Santana (2012) and Salaudeen et al. (2019). The FWO expression could be reduced as per the following: Henceforth, a linear relationship is obtained by plotting log (β) versus 1/T at an interval of 5% conversion (similar to the case of Friedman's). It is also irrelevant to consider the actual mechanism of degradation in the FWO which will not affect the Ea estimation. The third and last method used in this work is the Kissinger method which relies on the maximum degradation temperature of the material in the microbalance with respect to each β value. The slope of straight line obtained by fitting lnð β � T 2 m Þagainst (1/T m ) is used to estimate the E a values with respect to the maximum decomposition denoted as T m (Al-Salem et al. 2017). The Kissinger method is noted to be a model-free one which yields a single value E a similar to the previous isoconversion methods considered (Blaine and Kissinger 2012). Figure 1 depicts a summary of this works research content for the reader's consideration.

Thermal stability and thermogravimetric analysis results
Passenger vehicle (car) tyres constitutes three main elements which are rubbers (e.g., natural-NR, styrene, styrene-butadiene-SBR), carbon black (CB) and additives of both organic and inorganic origins. Tyre properties and characteristics can significantly impact their thermal (and subsequently, thermokinetic) behavior and yields of thermo-chemical conversion (TCC) products which are governed by the proximate and elemental analysis, particle size, origin and history of the material as documented in past works (Antoniou and Zabaniotou 2015;Hita et al. 2016;Ucar et al. 2005;Williams 2013;Williams and Bottrill 1995). The main elemental analysis properties and proximate ones of the materials used in this work are summarized in Tables S1-S2 as aforementioned (Al-Salem 2022). Table 2 shows the main thermal properties extracted from the TGA experiments with respect to each particle size (milling process) and condition (e.g., non-dried and dried). Figures S7-S12 show the TG thermograms of the materials under investigation herein. There was a noticed shift in the degradation curves to a higher temperature and subsequently an increase in both values of the T os and T max which was proportional to heating rate (β). This could be attributed to the heat transfer effect which diminishes with smaller particle size as noted in the examined dataset below and past works (Al-Salem et al. 2017). Unapumnuk et al. (2006) detected weight conversion shifting to a T max higher than 600°C, whilst ELT particles in the case of Antoniou and Zabaniotou (2015) stabilized in weight loss with a clear shift in thermograms reaching higher than 700°C. Rajkumar and Somasundaram (2022) reported the TG behavior of ELTs and noted a shift in the thermograms reaching a maximum of almost 650°C. Previous investigations on ELTs show that T os in pyrolysis experiments is within the range of 200 to 250°C depending on experimental conditions and particle size (Al-Salem 2020; Al-Salem, Lettieri, and Baeyens 2009; Antoniou and Zabaniotou 2015; Berrueco et al. 2005; Jansen, van der Walt, and Crouse 2022; Menares et al. 2020). By closely examining the data at hand resulting from our work, we can generally notice three temperature regions (weight loss curves) in the TG thermograms ( Figures S7-S12). The first temperature region starts from 100°C to 250°C, which can be attributed to the loss of moisture, extender oils and plasticizers/additives from the tyre particles (Antoniou and Zabaniotou 2013). The second region extends to about 350°C which is attributed to the degradation of the rubber fractions (NR and BR) in the tyres (Antoniou and Zabaniotou 2015). The last region of thermal degradation is between 350°C and 550°C which is related to the degradation of the SBR fraction. The slow weight loss that occurs past 550°C has been previously reported to be due to degradation of the cyclized/cross-linked polybutadiene residues and the cracking/volatilization of high-molecular weight residuals in the char matrix (Jansen, van der Walt, and Crouse 2022;Senneca, Salatino, and Chirone 1999). The effect of the heating rate (β) and the milling process was clearly evident on the produced dataset ( Table 2). The lower the β is, the lower the estimated T os estimated. Lower rates result in revealing the true mechanism of degradation and are recommended to be used as a baseline for comparative assessment between non-isothermal TG datasets (Vyazovkin et al. 2014). The reliability of kinetic data related to mass loss are strongly affected by the heating conditions and the experimental conditions due to sample mass, gas atmosphere, and heating programme (Vyazovkin et al. 2011). Higher heating rates on the other hand have been reported to result in a steep degradation curve that accelerates toward maximum end of set temperature due to the endothermic effect of the global pyrolysis reaction (Grieco, Bernardi, and Baldi 2008). This was also confirmed prior using various analytical kinetics methods showing that the heating rate influences the kinetics triplet by Torres-Sciancalepore et al. (2022). The drying of the samples resulted in an increase in the T os by some 30°C. However, cryomilling the ELTs samples showed a great effect in reducing the onset of the degradation curve by about 10°C. The smaller the particle (as in the cryomilled samples) the larger surface area they possess, which also results higher volatiles release with the pyrolysis reaction (Antoniou and Zabaniotou 2015). Past published literature has also documented various T os data with various particle sizes of ELTs and virgin rubbers. Antoniou and Zabaniotou (2015) conducted TGA experiments in dynamic mode on coarser ELTs particles of size approximately ranging between 15 and 20 mm. The T os was recorded at about 250°C with a clear shift in the curves, similar to our case in this work, between the β values of 5°C to 30°C min −1 . The work of Berrueco et al. (2005) in isothermal mode (300-500°C) with ELTs of 20 × 20 mm in size resulted in a start of decomposition at 200°C. Jansen, van der Walt, and Crouse (2022) subjected 1 to 5 mm truck tyres granules to N 2 pyrolysis in dynamic mode and showed that the onset of degradation was about 250-300°C depending on the value of β Tread, sidewall and inner-liner pieces of ELTs were subjected to dynamic TGA (5°C to 25°C min −1 ) by Shahi, Dwivedi, and Manjare (2022). The T os was recorded at about 250°C for the different materials with slight variations indicating that their size was large than 5 mm. The past literature survey conducted herein verifies that the there exists a clear correlation between the size (and shape) of ELTs in TGA and their thermal properties. This work has also pointed out this difference clearly and showed that (other than drying) but also no drying has minimal effect with milled sample using two different techniques. However, when drying is conducted; we can clearly notice that a variation in the sample's thermal properties.

Heat flow calorimetry and chromatographic analysis
DSC analysis of heat flow characteristics has been conducted in this work on three ELTs specimens under consideration. A vast body of work has previously employed DSC analysis to characterize melt peaks (melting point, T m ) and endothermic peaks to elicit information with regards to the materials tested. Napoli et al. (1997a) studied the isothermal behavior of ELTs and reported the mass loss to surpass 60 s and a peak at about 400°C. Similar techniques were also employed on different polymeric  Figure 2 shows the variation in heat flow analysis against the temperature for each specimen. At the start of the heating program, an endothermic process is witnessed in the form of a spike in the beginning of the curve which is attributed to the evaporation of industrial solvents and water reaching till about 100°C (Knapčíková, Herzog, and Oravec 2010). A large exothermic region is recorded between about 250 to 400°C and an endothermic region that terminates at 500°C (Figure 2). The endothermic energy could be attributed to the rupture of chemical bonds in the rubber matrix (Napoli et al. 1997b). The significant thermal events that all materials exhibit in this work also implies that they possess high crystalline regions in their matrix (Joohari and Giustozzi 2021). The type of sample, in terms of size and processing/milling condition, had a significant impact on the heat flow properties and calorimetry reported. It is true that two significant regions are detectable herein, the first is between 250 to 400°C and the second is between 400 to 500°C. The exothermic region for the original shavings resulted in an enthalpy of −48 J g −1 , whilst the cutting mill samples exhibited an enthalpy increase to −184 J g −1 . However, the cryomilled samples showed an enthalpy increase estimated at 139 J g −1 . The first exothermic region also terminates with the end of the first DTG peak for each material (Figure 2) which corresponds to the depolymerization of butadiene and vinilcyclohexene (Antoniou and Zabaniotou 2015). Hence, the rate of depolymerization is suspected to be controlled to a large extent by the size of the ELTs. This is due to the low heating rate that shows the mechanism of degradation, therefore the controlling step of the mechanism. Furthermore, the endothermic peak which corresponds to the T m varies significantly as well, whereby it was estimated to be 422°C (original shavings), 450°C (cutting mill samples) and 452°C (cryogenic samples) (Figure 2.e). The DTG curve also shows two distinct peaks that were previously noticed and discussed by Acevedo et al. (2015). The first which is recognized clearly at 375°C for the shavings (Figure 2.b) and cutting mill samples (Figure 2.d) corresponds to the presence of natural rubber and depolymerization/ degradation of butadiene and vinilcyclohexene compounds. The peak/dip is not significant and barely recognizable in comparison to the in the case of cyromilled samples (Figure 2.e). The second peak in the DTG is recognized clearly for all samples at about 450°C and is related to the dissociation of the butadiene rubber with styrene-butadiene rubber. Gao et al. (2022) showed that the maximum degradation between 450 to 500°C is attributed to oils and volatiles release in ELTs. This also corresponds to the experimental results witnessed in our work. Berrueco et al. (2005) studied the degressive thermal behavior of 20 × 20 mm tyre shavings and recorded under three distinctive peaks in the isothermal DTG curves between 300 to 600°C. These peaks were in the following temperature regions: 200°C to 325°C, 325°C to 400°C and 400°C to 500°C which correspond to the decomposition of the mixture of oils, moisture, plasticizers and other additives; and that the second and third peaks correspond to the decomposition of the main tyre components (e.g., natural rubber (NR), polybutadiene (BR) and polybutadiene-styrene). A similar behavior to our work presented in this article showing two distinct peaks was witnessed by Chen, Chen, and Tong (2001), Danon et al. (2015a;2015b), Rijo, Dias, and Wojnicki (2022). Figure 3 shows the major detected compounds in GCMS for the three types of specimens analyzed in this work. A great similarity in chemical compounds was detected between the original shavings of ELTs and cutting mill samples studied by examining the normalized area of the chemical compounds and judging from the retention time of the compounds which was similar in both cases (Figures 3.a-3.b; Figure 4). The cutting mill process facilitated the increase in styrene which increased from 5% to 22% (Figure 3.a). This could be the direct result from the smaller particles that facilitate depolymerization of SBR. Benzaldehyde showed a significant decrease from 31.3% to 22.7%. Furthermore, 1,3-diphenylpropane and 3-diphenyl-1-butene have also showed a slight decrease in comparison to original shavings.
Benzene, toluene, and styrene have all showed a significant increase in the cryomilled samples in comparison to the shavings (Figure 3.c and major chemical's retention times are depicted in Figure 5). The main rubber constituents of the ELTs (NR, SBR and BR) possess C-C, C=C and C-H bonds contained in either aliphatic or aromatic compounds. Since the required energy to dissociate C=C is higher than the other two in pyrolysis (Alkhatib et al. 2015), the thermal cracking is given priority to the compounds with C-C and C=C which milling has proven to aid in their formation Figures 4-5 This also facilitates findings by Zedler, Wang, and Formela (2022) whereby ground tyres is appropriately used for high adsorption capacity after functionlisation based on detected compounds. It is therefore concluded that the milling process, especially cryogenic milling (cryomill), can aid in characterizing and by association facilitating the production in aromatics more easily from ELTs when conducted on vast amount of feedstock in pyrolysis.

Degressive thermokinetics
In this section, we analyze the effect of the drying protocol used in the TGA experimental set-up on the estimated apparent activation energy (E a ) determined using a number of reaction kinetics methods previously depicted in Section 2 (Vyazovkin et al. 2014). The information extracted from the TG curves with respect to various β (5°C to 25°C min −1 ) was used accordingly. The differential based Friedman's curves are presented in Figures S13-S18 for the reader's consideration whereby we depict the plot of ln(dα/ dt) vs 1000/T, in order to obtain E a (in kJ mol −1 with a 5%) interval of conversion (%). We first examine the dataset results of the non-dried specimens where the maximum conversion rate (a) was estimated for the shavings as 70% and for both cutting mill and cryomill samples as 60% and 65%, respectively (Figure S13-S15). As for the dried specimens, the shavings showed a maximum conversion estimated as 60% (Figures S16) and the cutting mill and cryomill samples exhibited a maximum conversion of 55% and 65% ( Figure  S17-S18). The variation in Ea values with respect to the rate of conversion is represented in Figure 6. There was a large variation in the Friedman's results judging from the standard error (se) of the results especially at intervals above than α = 0.5. The se for the non-dried samples varied between 4 and 19 kJ mol −1 ; and for the dried samples between 8 and 24 kJ mol −1 . Past works also reported large variation in Friedman's results proportionally increasing with conversion rates (Mkhize et al. 2019). The variation in the Friedman's method also implies that the degradation reaction follows complex kinetics that the Friedman's method cannot represent (Vyazovkin et al. 2011), especially considering the fact that the method is mechanismindependent and estimate the entire average E a across the different stages of degradation (Al-Salem et al. 2017). Table 3 summarizes the E a estimated for the materials tested in this work. The non-dried samples showed the highest E a value when the cryomill samples were tested (181 kJ mol −1 ). This is attributed to the large surface area that needs to be exposed to the thermal conditions and subsequently affects the thermokientics estimation. The same behavior was also noted for the dried samples in comparison to the ELTs shavings (236 kJ mol −1 ).
In the case of the samples that were milled using the cutting mill, a decrease in E a was noted (53.13 kJ mol −1 ) in comparison to the original shavings of non-dried ELTs (94.25 kJ mol −1 ). The uniformity of the cutting mill samples is not comparable to the cryomilled ones, which could be the reason behind the drop in the estimated E a value (Table 3). Chen, Chen, and Tong (2001) tested car and truck ELTs powers after screening them with a 355 μm and 2 mm meshes, respectively. Friedman's method was applied on the TG dataset and resulted in E a equal to 147.64 kJ mol −1 (car ELTs) and 148 kJ mol −1 (truck ELTs). Four types of single component rubbers were studied in Danon et al. (2015a) work namely PI, SBR, NR, and BR. The analysis using the Friedman method yielded similar results of the PI and NR with the shoulders of the curves showing the highest E a values at around 400 kJ mol −1 with minimum estimations reaching 200 kJ mol −1 .
In order to estimate the E a of isoprene and DL-limonene, Mkhize et al. (2019) used the Friedman method on TGA curves for ELTs of approximate size (5 mm). The formation of isoprene was estimated at 141 kJ mol −1 which was relatively constant with respect to α values. The formation of DLlimonene on the other hand was estimated at 145 kJ mol −1 between values of α ranging from 0.1 to 0.5. The FWO method was also applied on the tested materials and plots of which could be found in Figures S19-S24. A smooth trend was observed amongst the study samples population especially amongst the milled samples with regression coefficients (r 2 ) exceeding 0.9. The FWO has also followed suit to the other isoconversion method used in this work (i.e., Friedman's) whereby the non-dried samples exhibited an E a of 78.43 kJ mol −1 (original shavings) and by comparison the cutting mill samples reduced in energy (54.48 kJ mol −1 ) and a notable increase was in the cryomill samples estimated as 212.73 kJ mol −1 (see Table 3). The drying of samples will not only facilitate the dissociation of the rubber matrix, but will also lead to a more uniform heat distribution which results in an incremental E a increase with smaller particles (see E a values in Table 3). The estimated E a values in this work are slightly lower than past ones using FWO method, which is expected as particle size varies significantly, as well as, experimental conditions (Cherop, Kiambi, and Musonge 2017). Future studies could extend from this work as well, in terms of co-processing ELTs with various other feedstock materials to report the effect on Ea in relation to various kinetic models. Such materials include petroleum vacuum residues (Ahmaruzzaman and Sharma 2007), commingled plastics (Biswas, Mohanty, and Sharma 2013), oils (Biswas and Sharma 2014) and biomass (Biswas and Sharma, 2021). Figure 7 shows the Kissinger plot for the dataset under consideration. The method relies on the maximum degradation temperature which accounts for maximum conversion for each material. It clearly noticed that both sets of data exhibit smoother lines in comparison with those obtained from the previous two methods considered. In the case of non-dried samples, the maximum E a was estimated for the cutting mill samples (147.35 kJ mol −1 ), whilst in the case of dried samples, the shavings were highest (242.89 kJ mol −1 ) ( Table 3). The Kissinger method is an isoconversional method that relies on an integration approach that typically produces straight and uniform lines since it relies on results before nominal reaction progress with high confidence (Mkhize et al. 2019). Rijo, Dias, and Wojnicki (2022) used to estimate Ea of ELTs (0.6-0.8 mm) using the Kissinger method. The results were in the range of 82 to 127 kJ mol −1 which is slightly lower than estimates in the current work as the sample size was higher in their work. Menares et al. (2020) determined the mechanism of ELTs degradation using Py-GCMS studies to reveal that it starts by degradation of main components of the tyre (e.g., SBR, NR, BR) in a two-way parallel reaction to limonene and isoprene from NR (i.e., polyisoprene). Xylene is produced from the limonene and a Diels-Alder reaction type controls the reversible reaction to xylene which also produces toluene and leads to benzene production. SBR on the other hand degrades and produce's styrene and ethylbenzene and contributes to toluene production too. SBR also produces 1,3 butadiene which leads to gases formation. Finally, BR degradation is the main contributor to cyclopetadiene formation. Judging from the results obtained in this work, we can come up with a number of major concluding highlights. The fact that styrene was estimated as 5% in the shavings of the ELTs and has increased in samples subjected to cutting mill (22%) and cryomill (38%). This clearly indicates that the milling process (proportionally to sample's particle size) facilitates the degradation of SBR. Generally, the drying for a period of 5 minutes which was conducted in past research (Danon et al. 2015a) has shown the it had an effect on the estimated E a values (Table 3). The drying has increased values of E a and was clearly noted with respect to particle size reduction using the Kissinger method (considered to be the most reliable as it takes into account the maximum degradation temperature). The drying could be a matter of great concern when cost analysis is conducted on samples treated in semi-pilot or pilot plant units of pyrolysis, which will lead to an increase in capital and operating costs of energy consumption.

Conclusion
End of life tyres (ELTs) were subjected to milling in a cutting mill and a cryomill, in order to, investigate the effect of conditioning on their thermogravimetric and thermokientic characteristics. The samples were subjected to pyrolysis in a dynamic TGA set-up under various heating rates between 5°C and 25°C min −1 . The experiments also involved drying of a period equal to 5 minutes, to be able to compare the data against non-dried samples, to further investigate the effect of drying on the results obtained. The thermograms obtained showed three distinct regions between 100°C and 250°C, 250°C to 350°C and an extended one till 550°C. These degradation regions within the thermograms can be directly linked to loss of volitiles and oils in the tyre grades, which on the other hand; could provide an insight as to when products can evolve in such reactions. The drying of the samples resulted in an increase of the T os by some 30°C. The drying protocol shows that future work could be optimized in energy use between drying requirements and prolonging degradation of ELTs. However, cryomilling the ELTs samples showed a great effect in reducing the onset of the degradation curve by about 10°C amongst the same set of experiments. The type of sample, in terms of size and processing/milling condition, had a significant impact on the heat flow properties and calorimetry reported. It is true that two significant regions are detectable herein, the first is between 250°C and 400°C and the second is between 400°C and 500°C. The exothermic region for the original shavings resulted in an enthalpy of −48 J g −1 , whilst the cutting mill samples exhibited an enthalpy increase to −184 J g −1 . However, the cryomilled samples showed an enthalpy increase estimated at 139 J g −1 . Drying of samples has also led to a notable increase in the estimated apparent activation energy estimated in this work using various method. It is easily envisaged that the kinetics derived from this work could be used by researchers in the future to design thermo-chemical conversion units that operate on pyrolysis and/or gasification which can accommodate the thermal characteristics of ELTs feedstock. The drying protocol used in this work could also be used in the future to design an upstream system for units that require pretreatment before pyrolysis. In addition, drying could be coupled with size reduction to target specific products based on the aforementioned. The work here also provides a perspective on adding grinding and milling, in addition to, additional drying stages prior to treating ELTs. All of which can be taken into account when scaling up is considered for treating rubber waste.

BR
Polybutadiene Rubber ICTAC International Confederation for Thermal Analysis and Calorimetry m Initial mass of rubber samples (g) m f Final mass of rubber sample (g) m o Mass of rubber at specific reaction time (g) N 2 Nitrogen Gas, NB, Natural Rubber SBR Styrene Butadiene Rubber T 50% Temperature of 50% Loss (

Acknowledgment
The lead author/project leader would like to thank the Kuwait Foundation for the Advancement of Sciences (KFAS) and the Kuwait Institute for Scientific Research (KISR) for funding and supporting this research project through the Grant for Project EM085C (PN17-44SC-03). The lead author would also like to thank Kuwait Municipality (KM) for their help and support to the work conducted in this research. The Project Leader would also like to dedicate this report to Mr. Majed Al-Wadi who has retired from service as a Principal Senior Research Technician after a fruitful and prosperous 33 years career at KISR ending his work with duties assigned to this project.

Disclosure statement
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.

Funding
The work was supported by the KFAS [EM085C].

Data Availability Statement
Raw data were generated at KISR. Derived data supporting the findings of this study are available from the corresponding author (S.M.S.) on request.

CRediT authorship statement
S.M. Al-Salem; Conceptualization; data analysis; fund acquisition; initial and final draft preparation, H.J. Karam; Experimental work; data analysis.