Synergistic effect of mica , glass frit and melamine cyanurate for improving fire resistance of styrene-butadiene rubber composites destined for ceramizable coatings

Synergistic effect of different fillers is widely utilized in polymer technology. The combination of various types of fillers is used to improve various properties of polymer composites. In this paper a synergistic effect of flame retardants was tested for the improvement of ceramizable composites performance. The composites were based of styrene-butadiene rubber (SBR) used as polymer matrix. Three different types of flame retardants were tested for synergistic effect: Mica (phlogopite) high aspect-ratio platelets along with low softening point temperature glass frit (featuring ceramization effect) and melamine cyanurate, a commonly used flame retardant promoting carbonaceous char. In order to characterize the properties of the composites, combustibility, viscoelastic properties and mechanical properties before and after ceramization were tested. The results obtained show that the synergistic effect of ceramization promoting fillers and melamine cyanurate is especially visible with respect to the flame retardant properties resulting in a significant improvement of fire resistance of the composites.


Introduction
Ceramizable (ceramifiable) polymer composites were originally designed and developed for cable coatings ensuring integration of an electric installation in case of a fire-hazard [1][2][3].Such approach resulted in a significant enhancement in evacuation techniques and fire-fighting systems.For example, allowing for the development of fireproof elevators ensuring effective evacuation possibilities, especially for disabled or elderly people [4].Moreover, in recent years the area of ceramizable composites utilization expanded covering the fireproof window sealing system [5], coatings for steel construction elements [6], and innovative ablative materials for spacecraft applications [7], starting to compete with much more popular now intumescent coatings [8].
Ceramizable polymer composites are a type of smart materials changing their structure and properties when exposed to fire and elevated temperature.They exhibit high elasticity and easy processing before ceramization and transform into a continuous and stiff ceramic structure after ceramization.From a morphological point of view they are dispersion composites consisting of polymer matrix and a mix of dispersed fillers that initiate the ceramization process.The polymer matrix burns and decomposes at elevated temperatures, while the fillers interact forming a continuous ceramic structure.
There are several mechanisms of ceramization: (1) addition of a low softening point temperature glass frit that acts as a binding agent for other thermally stable fillers [9][10][11]; (2) sintering of filler particles accompanied by silicone rubber matrix covalently bonded to their surface [12]; (3) in-situ formation of calcium silicates from calcium-based fillers and the silica formed after silicone matrix burning [1]; (4) silicon oxycarbide phase formation from cross-linking of silicone rubber adsorbed on nano-filler surface or catalyzed by platinum compounds [13]; (5) in-situ ceramic structure formation from reactions between the products of thermal decomposition of silicone rubber, ammonium polyphosphate, aluminum hydroxide, and mica [14]; and (6) controlled arrangement of mineral fillers inside the polymer matrix [15].A specific type of secondary filler may be used in order to improve the ceramic structure's properties, for example, triggering the formation of a nano-porous phase [16] or enhancing its compression strength [17].
Silicone rubber was originally chosen as a ceramizable composites matrix because amorphous silica is the product of its burning, which improves the properties of the ceramic phase.However, silicone rubber exhibits also some limitations like relatively low mechanical properties, high price, and low filler loading.Therefore, alternative polymer matrices are tested for ceramizable composite applications, such as: acrylonitrile-butadiene rubber (NBR) [18], ethylene-propylene-diene rubber (EPDM) [19][20][21], poly(vinyl acetate) [22], polyurethane [23], polyester [24], polyethylene [25], and boron phenolic resin [26][27][28].However, their applicability as ceramizable composites remains limited.The most promising replacement for silicone rubber nowadays seems to be the poly(ethylene-co-vinyl acetate) copolymer (EVA), ensuring good mechanical properties both before and after ceramization [29][30][31][32].The positive effects of EVA application as a matrix for ceramizable composites resulted also in using it as a blend with silicone rubber [33].However, the use of silicone rubber alone is still very prospective [34,35].In our previous work we proved that the application of styrene-butadiene rubber (SBR) as a polymer matrix for ceramizable composites is also very promising [36].Based on the available literature [37][38][39], one can state that thermal stability of SBR is comparable to that of silicone rubber [40,41].Moreover, in the presence of oxygen, SBR degrades with two-stage kinetics [38], first occurring in the temperature range of 390-520 • C and subsequently between 620 and 720 • C. The first stage most probably corresponds to the thermal degradation of linear rubber macromolecules and the simultaneous cross-linking (maybe even carbonization) leading to the formation of a thermally stable residue that decomposes thermooxidatively in the second stage.The highly cross-linked structure formed during thermal degradation plays most likely a significant role in maintaining a coherent, self-supporting structure of the ceramizable composites in the temperature range where the elastic SBR matrix is gone but the new ceramic structure is not fully formed yet [36].For this reason, SBR rubbers emerge as very prospective matrix materials for ceramizable composites.
In the current work we test the effect of the melamine cyanurate (MCA) addition on the properties of SBR-based ceramizable composites.We aimed to investigate the potential synergistic effects of MCA, mica platelets, and glass frit promoting ceramization on the flame retardancy performance of composites.MCA has proven itself an effective rubber flame retardant [42,43], and has been already successfully applied for enhancing flame retardancy of EVA-based ceramizable composites [31].Furthermore, the combination of glass frit and mica for SBR-based composites showed an outstanding ceramization performance [36].

Preparation of the Samples
The composite mixes (Table 1) were prepared using a laboratory two-roll mill (the rolls length 200 mm; diameter 150 mm; Bridge, UK), working with the rotation speed of the slower roll of 18 rpm (revolutions per minute) and the faster roll of 20 rpm (friction: 1.1).Kinetics of vulcanization of the composite mixes were tested by means of rheometer Alpha Technologies MDR2000 according to ISO 37:1994 standard [44] (temperature: 160 • C, frequency: 1.667 Hz, strain: 0.5 • ).According to the results obtained) the samples were shaped and vulcanized in steel molds by a laboratory press at 160 • C and under 10 MPa of pressure.
Mechanical properties of the vulcanizates were tested by Zwick/Roell 1435 testing machine-before and after ceramization, and Zwick/Roell hardness tester (Ulm, Germany)-before ceramization.
Combustibility of composite vulcanizates was determined by a cone calorimeter (Fire Testing Technology Ltd., East Grinstead, UK).Samples with dimensions of 100 mm × 100 mm × 2 mm were placed horizontally in relation to the IR heating source of 35 kW/m 2 .The oxygen index test was performed on a home-made device in accordance with the ISO 4589-2:2017 standard [46].Vertical fire spread test was performed in accordance to UL-94 standard [47].Thermogravimetric analysis was done under synthetic air or nitrogen atmosphere with heating rate of 20 Micromorphology of the composites was examined before and after ceramization by means of scanning electron microscopy (SEM) Jeol JSM-6400 (JEOL Ltd., Tokyo, Japan) with Noran energy-dispersive X-ray spectroscope unit (EDS) (Thermo Fisher Scientific, Waltham, MA, USA).Sample cross sections were prepared by crushing in liquid nitrogen and gold sputtered directly before the measurements.

Viscoelastic Behavior and Vulcanization of Composites
The complex dynamic viscosity η* which is measured by an oscillating rheometer is analogous to the viscosity η app measured by a capillary rheometer [48][49][50][51].The complex dynamic viscosity η* is calculated from parameters obtained from rubber process analyzer (RPA) as follows: The complex shear modulus (G*) relation to storage shear modulus (G') and loss shear modulus (G") is presented by the Equation (1): The complex dynamic viscosity (η*) relation to complex shear modulus (G*) is given by Equation ( 2): where ω-frequency (Hz).Dynamic viscosity values both for low and high strain grew with the increasing quantity of mineral fillers (mica and fluxing agent).The addition of a fluxing agent had small impact even for low frequencies (Cyan_1 and Cyan_3; Figure 1 and Table 2).Mica used as filler with platelet-shape particles changed the viscosity of the composites to the highest extent.It is especially visible for the composite Cyan_5 where the general content of additives is the same as for the composite Cyan_6.For which 25 phr of mica and 25 phr of fluxing agent was exchanged to melamine cyanurate.However, the changes in the mix dynamic viscosity was almost two times (low strain) and three times (high strain) higher.Table 2. Viscoelastic properties of uncured composites at chosen frequencies: storage shear modulus (G'), loss shear modulus (G"), and complex dynamic viscosity (η*).What is surprising is the large difference between Cyan_4 and Cyan_5.The composition of both composites differed only by 25 phr of melamine cyanurate, and the dynamic viscosity changed from 7 to 18 MPa•s (low strain, 0.628 rad/s), and from 1 to 3 MPa•s (high strain, 0.628 rad/s).Given the insignificant/slight influence of MCA alone on mix viscosity (Cyan_1) it is clearly a synergistic effect between the mineral fillers and MCA.Probably this combination of the mineral fillers and melamine cyanurate results in a strong secondary structure formation, in which the mineral particles are connected with low-molecular MCA.It is clear that MCA itself is not able to form a secondary structure in the SBR matrix (Figure 2).However, in the presence of additional mineral fillers it behaves like inter-particular low molecular binder resulting in a strong Payne effect.It is especially visible for composites Cyan_5 and Cyan_6.When the amount of fillers was reduced, the viscosity and G' values dropped significantly.It is noticeable in the literature that application of low-molecular additives into rubber matrices may alter their viscoelastic properties to a significant extend [52].The viscosity of the mixes is a crucial factor during their processing, especially extrusion, which is commonly used for cable production.If the viscosity of rubber mixes is too high, their processing may be much more difficult or even impossible.Therefore, knowledge regarding the ability of a mica/glass frit/MCA system to form a strong secondary structure is of great importance.

Composite
The value of storage shear modulus G' as a function of growing strain shows how the secondary structure formed by fillers is destroyed, thus reducing the viscosity of the system.This effect is independent from a polymer matrix properties, but is caused by the specific filler-filler interactions, whose range is limited.Melamine cyanurate, both with mica and a fluxing agent, creates a structure with almost the same interactions (Figure 2).The highest values in dynamic viscosity for the Cyan_5 composite can be the result of the strongest interactions, which are created between filler particles connected by the low-molecular MCA.This can be explained by the chemical nature of the fillers-all three of them exhibit highly polar characteristics in opposition to the SBR matrix.Therefore, they are more likely to interact with each other resulting in a strong filler-filler secondary structure.In each case the addition of all three types of fillers causes stronger interactions than in the case when one of the fillers is not used.
The incorporation of fillers into the e-SBR rubber matrix resulted in a significant decrease of scorch time and optimum vulcanization time (Table 3 and Figure 3 and Figures S1-S6).The chemical structure of the melamine, containing three amine groups, indicates its possible vulcanization-acceleration feature.The application of only 50 phr of MCA resulted in a scorch time value of 5 30".Surprisingly, the incorporation of additional mineral fillers (mica, glass frit) decreased this value very significantly.This is most probably the effect of a better distribution of MCA in the rubber matrix in the presence of additional mineral fillers.The incorporation of mineral fillers increases the shear forces during mixing resulting in better dispersion of MCA.Moreover, it is possible that some MCA molecules adsorb on the filler surface, which additionally enhances their dispersion and distribution in the e-SBR rubber matrix.This phenomenon may pose a threat on the processability of the composite mixes causing premature vulcanization.Increase in torque value stabilizes on a plateau for all composites showing no signs of reversion during the one hour of testing.The addition of fillers does not influence the glass transition temperature (T g ), which fully depends on the SBR matrix (Figure 4).Composites' T g values oscillate around -48 • C, which is equal to unfilled and uncured SBR given by its producer.Apparently large particles of mica, glass frit, and MCA do not interfere segmental macromolecular movement.This was expected because of particle size and their poor interfacial compatibility with the rubber matrix.All the fillers exhibit polar surface properties, whereas the rubber is highly non-polar.Therefore there is no factor that could constrain the segmental movement of the SBR macromolecules.

Mechanical Properties of Composites before Ceramization
The effect of forming strong interactions between filler particles via MCA acting as a binder is also visible in mechanical properties of vulcanized composites (Table 4).For the composites, which were characterized by the three highest values of storage shear modulus G', the elongation at break never reached 200%.Thus, composites elasticity was visibly reduced (except of Cyan_3) in comparison to the ceramizable composites we tested previously [36].However, Cyan_5 which exhibits the highest G' and tensile strength values shows higher E b than Cyan_4 and Cyan_6.This is due to the most effective secondary structure formation by an optimal ratio of the incorporated fillers: 200/100/50 for mica/glass frit/MCA respectively.For Cyan_4 the amount of MCA is too low to act effectively as a low-molecular binder for the mica and glass frit.Whereas, for Cyan_6 the amount of the mineral fillers was too low with respect to MCA, which alone was not able to form a secondary reinforcing structure as shown by the Payne effect measurements (Figure 2).Therefore, the largest number of strongly connected filler particles was present in the Cyan_5 composite.Consequently, the value of Cyan_5 tear resistance reached 27.2 N/mm, which is the highest value for all the tested composites.The tensile strength of composites oscillated between 4.2 and 6.4 MPa, which is comparable to our previous results [36] and allows successful industrial applications.The hardness of composites was most heavily impacted by the addition of mica.The addition of as little as 100 phr of other fillers to the composite, which already had 200 phr of mica did not affect its hardness, whereas the addition of fluxing agent affected the value of hardness but only when mica had not been already added.This is due to the fact that its stiff-platelet particulate structure increases the stiffness of composites significantly.In general, incorporation of fillers resulted in higher tensile strength and hardness of the composites, whereas elongation at break is reduced in comparison to the neat SBR vulcanizate [36].

Micromorphology of Composites before Ceramization
The micromorphology of samples investigated via SEM technique is presented in Figure 5. Incorporation of fillers resulted in significant changes in the samples' micromorphology.The glass frit and mica were particular mineral fillers exhibiting high value of average particle size.Especially the glass frit particles, which often have a diameter >10 µm (Figure 5f), which influences negatively the mechanical and dynamic properties of rubbers.However, decreasing the size of glass frits is a challenging task due to their amorphous nature and low softening point temperature.They are prone to crushing as well as sintering in the same time during milling.Therefore, the smallest fraction of commercially available glass frits is often <63 µm.On the other hand, mica particles exhibit platelet structure of high aspect ratio (Figure 5d), whose incorporation results in formation of anisotropic, layered morphology during the vulcanization mixes in the pressurized mold (Figure 5c).This seems to be very beneficial from the point of view of fireproof performance of the coating by limiting heat and oxygen transfer into the bulk of the composite as well as flammable products of the rubber matrix degradation into the burning zone.
Due to highly polar nature of MCA its dispersion and distribution was very poor in a non-polar SBR matrix (Figure 5a).Furthermore, the wettability of the MCA agglomerates by the rubber macromolecules was very low (Figure 5b), and it was not improved by incorporation of the glass frit (Figure 5g).Even though the viscosity of the mixes increased after the addition of glass frit (Figure 1 Coatings 2019, 9, 170 and Table 2), the increase was not high enough to improve the dispersion and reduce the size of MCA agglomerates.On the other hand, adding mica platelets improved dispersion of MCA to a significant extent (Figure 5e).Firstly, because the presence of mica increased the viscosity to a much higher extent than the glass frit.Secondly, the mica particles are much finer and harder so they were able to defragment the MCA agglomerates resulting in a very homogenous distribution of MCA, which is confirmed by EDS nitrogen mapping (Figure 6a,e).
In general the microstructure of the composites was heterogeneous due to large size of the mineral filler (mica, glass frit) particles.However, adding mica resulted in very good MCA dispersion and distribution.
The composites containing all three fillers (Cyan_4, Cyan_5, and Cyan_6) exhibited very similar micromorphology (Figure 5h-m).Therefore, we decided to conduct EDS mapping on Cyan_5 sample as an example of elemental distribution (Figure 6).The glass frit applied consisted of a high quantity of silicon and sodium, and therefore its particles were easily detected by the EDS technique in the composite microstructure.EDS measurements confirmed results from the SEM analysis showing the composites' heterogeneous structure.Especially, silicon (Figure 6h) and sodium (Figure 6f) maps indicated the position of glass frit particles.Nitrogen map (Figure 6e) confirmed good dispersion and distribution of MCA in the SBR matrix when mica was added.Magnesium (Figure 6d) and aluminum (Figure 6i) presence maps showed the location of mica particles indicating good dispersion and distribution of its particles in the composite.

Flame Retardancy
The synergistic effect of ceramization and melamine cyanurate was visible especially in flammability properties of the tested composites (Figure 7 and Table 5).It was only the addition of mica or fluxing agent to the composition containing MCA that resulted in a decrease in THR from 50.2 to 16.1 or 38.6 mJ/m 2 , respectively.For almost every combustibility parameter, the simultaneous effect of ceramization and melamine cyanurate was better than for each of them separately.Especially, the combination of MCA with mica resulted in significant improvement of flame retardancy of composites (Cyan_2) giving the lowest values of heat release rate, FIGRA (HRRp/tHRR ratio), and MAHRE (cumulative heat emission divided by time) parameters.This is due to the synergistic effect of carbonaceous char formation from MCA and large aspect ratio mica platelets, which can significantly improve the barrier properties of the char formed.This is most likely caused by the very homogenous dispersion and distribution of MCA in the presence of mica particles (Figure 5).Therefore, decreasing the transport of flammable hydrocarbon compounds formed from the thermal degradation of SBR rubber into the flame zone.Application of only glass frit with MCA did not improve fire resistance to a high extent, due to the fact that the frit does not promote effective ceramization alone and MCA dispersion is still poor.Time to ignition of the composites remained on the same level or increased from 89 to 222 s (Cyan_6), which was a significant improvement especially in comparison to the neat SBR vulcanizate [36].Similarly, the flameout time exhibited very high values, which meant that the burning rate of the composites significantly decreased by the flame retardant performance of the mica/glass frit/MCA system.Mass loss decreased with the growing content of fillers that promote ceramization.The oxygen index value was very high in comparison to flame retarded SBR composites described in the recent literature [53,54], and reached 37.0% or 37.5% for the composites without mica, and exceeded 37.5% for composites with mica (37.5% is the limit value for the home-made apparatus on which the tests were carried out).The high combustibility parameter values of the studied composites placed them among other novel flame retardant systems for rubber application [55].Table 5. Flammability parameters: time to ignition (t i ), time to flameout (t o ), heat release rate peak (HRR p ), heat release rate mean value (HRR m ), time to HRR p (tHRR), HRR p /tHRR ratio (FIGRA), total heat released (THR), effective heat of combustion peak (EHC p ), effective heat of combustion mean value (EHC m ), mass loss rate peak (MLR p ), mass loss rate mean value (MLR m ), cumulative heat emission divided by time (MARHE), mass loss (m l ), oxygen index (OI), and UL 94 fire test.For evaluating the flammability of the composites, the most important indicator is the FIGRA parameter value.This parameter is the value of Heat Release Rate peak divided by the time to reach this peak.Based on this parameter it is easy to compare the general flammability of composites, which depends not only on the heat generated during burning, but also on the time when this heat was emitted.The lower the value of this parameter the better fire resistance of a composite.Analyzing this parameter, the synergistic effect of ceramization and MCA-based char formation on flame retardancy of the composites were observed.For the composite containing only melamine cyanurate (Cyan_1), the value of this parameter was 1.34, the composite containing 200 phr of mica and 100 phr of fluxing agent-1.05.In each case, when melamine cyanurate and the full ceramization system were used, the value for this parameter was below 0.5: Cyan_4-0.39,Cyan_5-0.24,and Cyan_6-0.43.The presence of fluxing agent reduced the thermal properties (Cyan_2-0.19compare to Cyan_3-1.05)but not to a high extent.Nevertheless, this filler is necessary for creating strong ceramic structures during ceramization, which will also work as thermal and mechanical stress barrier.The combustibility parameter values were significantly improved when MCA was applied to the ceramizable compositions in comparison to the e-SBR based composites we studied previously [36] that did not contain MCA.

Thermal Stability
Thermal stability of the composites under oxidative (synthetic air) or inert (nitrogen) atmospheres is presented in Figure 8 and Table 6.Table 6.Thermal stability parameters of the composites: temperature of 5% mass loss (T 5 ), maximum rate of first (DRT max1 ), second (DRT max2 ), and third (DRT max3 ) step of degradation and their temperatures respectively (T max1 , T max2 , and T max3 ), residue after thermal degradation.Addition of mica platelets improves thermal stability of the composites (Table 6).The beginning of thermal degradation temperature (T 5 ) reached the highest value for the composite filled with only mica and MCA (Cyan_2).This was most probably due to the layered structure of mica platelets and anisotropic morphology of the samples that reduce heat flow (mica platelets arranged in parallel can act as mirrors reflecting the infrared heat rays) into the bulk of composites and oxygen penetration.Glass frit decreases thermal stability most likely by increasing heat conductivity.Moreover, the spherical particles of the glass frit might cause formation of voids filled with air during mixing.This could increase the rate of in-bulk thermal oxidation of the SBR matrix.

Combustibility
The samples tested under nitrogen atmosphere exhibited double step thermal degradation kinetics (Figure 8c).This is most probably due to the different decomposition kinetics of SBR and MCA.After incorporating mineral fillers, the first step moves to lower temperatures (Table 6).This might be a result of a catalytic effect of the mineral fillers surface promoting the thermal decomposition of rubber.At the same time, the second step of the decomposition took place under similar temperatures for all composites, suggesting that it originated from MCA decomposition.
The kinetics changed with changes in the test atmosphere to oxidative synthetic air (Figure 8b,d).An extra third step of thermal decomposition appeared for the samples measured under synthetic air at higher range of temperatures.This is most likely a result of residual carbon char burning formed from MCA.This phenomenon is confirmed by the literature, in which the decomposition of MCA filled polymers under nitrogen proceeds without the final third step [56].With the addition of mineral fillers, the temperature of the maximum decomposition of the third step of degradation shifted to lower temperatures.This is related to the amount of carbonized char formed during the thermooxidative decomposition: low-Cyan_3, Cyan_4 and Cyan_6 or high-Cyan_1, Cyan_2, and Cyan_5 (Figure 8b).

Properties of Composites after Ceramization
The composite containing only melamine cyanurate (Cyan_1) without fillers, which promote ceramization, in every case failed to create any ceramic structure, which was expected.The rest of the composites formed ceramic structures during heating but only the sample Cyan_3 did not change its shape much after ceramization to 1100 • C and could be tested for its compression strength.
The presence of both mica and fluxing agent was necessary to create a strong enough ceramic structure (Figure 9 and Table 7).The addition of mica (Cyan_2) or fluxing agent (Cyan_3) separately to the composite mix also resulted in a ceramic structure but with much worse compression strength properties.The application of melamine cyanurate along with mica and glass frit to the ceramizable compositions did not influence the mechanical properties of the composites after ceramization.In the composites with changing content of MCA (25-100 phr; Cyan_4-Cyan_6) the compression strength changed only within the limits of statistical error.Although it may seem that the error is very high, the samples' morphology after ceramization was far from homogenous in macro scale (Figure 10).Similar to random fire conditions, heating up relatively big samples containing many dispersed fillers results in their deformation due to several factors, such as: the initial inhomogeneity of their bulk structure; internal gas formation rate from degrading polymer matrix; carbonization of the polymer and MCA; thermal expansion of the mineral fillers; direction of the gas convection flow formed by the flame direction, etc.The appearance of composites after ceramization (Figure 10) for those containing mica was almost the same.The Cyan_1 composite evaporated during the heat treatment.The composite in which the glass frit was not used (Cyan_2), had many visible cracks being marks of a relatively weak inter-particulate bonding.There was no binding material that was able to effectively connect particles of mica during heating.The composites Cyan_4-Cyan_6 looked almost the same, which was also mirrored in their compression strength.For these samples, a stiff, durable and porous ceramic structure was observed.The sample Cyan_3 differed from the other samples since the fluxing agent during heat treatment connected only to the polymer matrix decomposition products accompanied with MCA carbonization residues, which resulted in the black color of this sample.However, the fact that Cyan_3 could withstand the severe ceramization conditions with only the carbonaceous char formed from MCA, SBR, and glass frit shows a high potential of these ingredients for ceramizable composites manufacturing.

Micromorphology of Composites after Ceramization
In order to examine more in depth the effects of ceramization on the composites micromorphology, the Cyan_5 sample was chosen as an example.The mechanical properties of Cyan_4, Cyan_5, and Cyan_6 after ceramization were similar as was their micromorphology.Therefore, Cyan_5 should be a good example for the investigation (Figures 11 and 12).SEM photographs of Cyan_5 sample ceramized in mild conditions (950 • C) revealed a solid and continuous microstructure (Figure 11) responsible for the good mechanical durability under compression test (Figure 9).However, the ceramic structure obtained was not homogenous on a micro scale level.Clear phase separation can be distinguished, especially under higher magnification (Figure 11b).This shows that the interphase adhesive forces are high resulting in good mechanical properties of the ceramic structure.For further phase analysis EDS mapping was performed (Figure 12).
EDS elemental mapping revealed the composition of the interphase structure (Figure 12a).Magnesium-rich islands clearly indicated (Figure 12b) the area where mica platelets are located on the surface of the sample's cross-section.High sodium concentration (Figure 12c) accompanied with the lack of magnesium and aluminum showed the area rich in softened glass frit in the upper part of the middle area of the photograph.Elemental analysis confirmed the non-homogeneity of the ceramic structure on a micro scale level.Regardless, high interfacial interactions between mica and glass frit resulted in high mechanical performance of the ceramized composite.

Conclusions
The addition of melamine cyanurate flame retardant improves flame retardant properties of ceramizable SBR-based composites.The synergistic effect is particularly visible for the fire resistance capacity of the composites measured by cone calorimetry, in which case, parameter improvement can reach six times better properties.Rheological properties of the mixes are also significantly affected by the synergistic effect of mica/glass frit/MCA compositions.The mineral filler particles are connected by MCA particles acting as binders form a very strong secondary structure in the rubber matrix.However, this requires an optimum components ratio and can easily be changed by decreasing the amount of MCA or mineral fillers.SEM-EDS analysis showed that the micromorphology of the composites is heterogeneous.However, incorporation of mica platelets greatly improves MCA dispersion and distribution in the SBR matrix.Mechanical strength properties of the ceramizable composites vary in a significant range, however, the changes are not very significant from the real application point of view.The incorporation of different amounts of melamine cyanurate does not change compression strength of the composites after ceramization simultaneously increasing their fire resistance.

Figure 2 .
Figure 2. Strain amplitude dependence of storage shear modulus G' for cured ceramizable composites.

Figure 3 .
Figure 3. Vulcanization kinetics of composite mixes measured 24 h after preparation.

Figure 4 .
Figure 4. DSC curves of the composites.

Figure 9 .
Figure 9. Compression strength of the ceramized composites studied.

Table 1 .
Composition (in phr-parts per hundred parts of rubber) of the ceramizable composites mixes.

Table 3 .
Vulcanization parameters of the ceramizable composite mixes.

Table 7 .
Compression strength of the ceramized composites studied.