Preparation and Thermo-Mechanical Characteristics of Composites Based on Epoxy Resin with Kaolinite and Clinoptilolite

Herein the synthesis, characterization, and study of spectroscopic, thermal, and thermo-mechanical properties of polymeric composites are presented. The composites were obtained in special molds (8 × 10 cm) based on the commercially available epoxy resin Epidian® 601 cross-linked by 10% w/w triethylenetetramine (TETA). To improve the thermal and mechanical properties of the synthetic epoxy resins, natural fillers in the form of minerals from the silicate cluster kaolinite (KA) or clinoptilolite (CL) were added to the composites. The structures of the materials obtained were confirmed by attenuated total reflectance-Fourier transform infrared spectroscopy (ATR/FTIR). The thermal properties of the resins were investigated by differential scanning calorimetry (DSC) and dynamic-mechanical analysis (DMA) in an inert atmosphere. The hardness of the crosslinked products was determined using the Shore D method. Moreover, strength tests were performed on the 3PB (three-point bending) specimen, with the analysis of tensile strains conducted using the Digital Image Correlation (DIC) technique.


Introduction
Epoxy compounds have been known since the late 19th century. Their development has progressed rapidly during this time. Today, epoxy-based plastics are used in many areas of human activity. Initially, they were substitutes for scarce natural materials and later became excellent replacements [1,2]. They are on a par with materials such as wood, metals, glass and ceramics, and many plastics surpass natural materials in their properties. Epoxy resins are brittle solids or viscous liquids that contain epoxy groups in their structure, which are capable of polyreactions that transform them into cross-linked, insoluble and infusible polymers. Cured resins are also referred to as epoxy resins, although they no longer contain epoxy groups [3][4][5][6].
The curing process is accompanied by fundamental changes in the physical properties of the resin-curing agent system. The curing reactions are exothermic and the heat released during the reaction process increases the temperature of the system in question so that the reaction between resin and hardener is accelerated. An undesirable temperature rise can be counteracted by adding the same amount of filler to the resin, which absorbs the heat given off. It is also possible to reduce the amount of hardener or to use a hardener with retarding additives [7][8][9][10][11].
Crosslinked epoxy resins are characterized by high abrasion, scratch and impact resistance. Other features are their compressive, crushing and tensile strength and high and next put into the oven (for 10 min at 50 • C). The beaker contents were poured into glass molds (10 mm × 8 mm × 3 mm) and polymerized for 24 h at room temperature. The samples were heated at 70 • C for 60 min after being tested. Figure 1 presents the proposed scheme of the composite structure. was obtained. Then the filler (kaolinite or clinoptilolite) was added in amounts of 1, 5, and 97 10 wt% to the weight of the epoxy resin + curing agent. The whole content was stirred and 98 next put into the oven (for 10 min at 50 °C ). The beaker contents were poured into glass 99 molds (10 mm × 8 mm × 3 mm) and polymerized for 24 h at room temperature. The sam-100 ples were heated at 70 °C for 60 min after being tested. Figure 1 presents the proposed 101 scheme of the composite structure.  .650 A total of 7 plates with dimensions of approximately 50 mm × 90 mm were desig-106 nated for bending tests, labeled and sequentially numbered as follows: CL5; KA1; CL10; 107 0 (reference material); KA5; CL1; KA10. 108 Complete sets of three or four samples with dimensions of approximately 80 mm × 109 12 mm × 3 mm were cut from the designated materials. Each sample was measured in 110 terms of cross-section before testing. When four samples were cut for analysis, the three 111 with the highest bending strength were selected. 112 114 The FTIR spectra were developed by applying attenuated total (internal) reflection 115 (ATR/FTIR) with the use of a Bruker TENSOR 27 spectrophotometer (Ettlingen, Ger-116 many), equipped with a PIKE measuring cell which features crystalline diamond embed-117 A total of 7 plates with dimensions of approximately 50 mm × 90 mm were designated for bending tests, labeled and sequentially numbered as follows: CL5; KA1; CL10; 0 (reference material); KA5; CL1; KA10.

Measurements Methods
Complete sets of three or four samples with dimensions of approximately 80 mm × 12 mm × 3 mm were cut from the designated materials. Each sample was measured in terms of cross-section before testing. When four samples were cut for analysis, the three with the highest bending strength were selected.

Fourier Transform Infrared Spectroscopy (FTIR)
The FTIR spectra were developed by applying attenuated total (internal) reflection (ATR/FTIR) with the use of a Bruker TENSOR 27 spectrophotometer (Ettlingen, Germany), equipped with a PIKE measuring cell which features crystalline diamond embedded in zinc selenide. The FTIR spectra were recorded within the range of 4000 to 600 cm −1 , with 64 scans per sample, at a resolution of 2 cm −1 in the absorption mode. and was first cooled to an isotherm for 3 min at −100 • C and then heated up to a maximum temperature of 200 • C, next cooled to −100 • C and then heated to 200 • C. The scans were performed at the heating/cooling rate of 10 • C/min under an argon atmosphere (gas flow = 30 cm 3 /min). All DSC measurements were taken in aluminum pans with a pierced lid (a mass of 40 ± 1 mg). As a reference, an empty aluminum crucible was applied. The reported transitions were taken from the first and second heating scans. Glass-transition temperatures (T gs ) for the polymer samples were taken as the inflection point on the curves of the heat-capacity changes. Curing temperatures (T curs ) were read at exothermic-peak maxima.
Dynamic Mechanical Analysis (DMA) DMA of composites was performed in tensile mode using a DMA Q800 Analyzer TA Instruments (New Castle, DE, USA). Calibration was performed as per the manufacturer's recommendations included in Advantage Software, version 5.5.24 (TA Instruments, New Castle, DE, USA). The experiments were carried out on rectangular samples of dimensions close to 3 mm thick, 10 mm wide, and 60 mm long. Experimental conditions employed were a frequency of 1 Hz and amplitude of 10 µm with the scanning air temperature range from −100 • C to 127 • C and a temperature ramp of 3 • C/min. The variations of storage modulus (E ), loss modulus (E ), and tangent delta (tan δ) versus temperature were determined.

Mechanical Properties
The hardness of the composites was measured by the Shore D method on a Zwick 7206/H04 hardness tester (Ulm, Germany). The readings were taken after 15 s at a temperature of 23 • C [28]. Bending strength was measured using a Zwick 2.5 kN static machine. Pairs of images to calculate major and minor strain directions were recorded by an ARAMIS SRX, Zeiss (Jena, Germany) DIC system. The cross-sectional area of the sample along with the support span and force information was used to determine the bending strength according to Formula (1): where: σ-bending stress, F-maximum force, l-support span, b-sample width, hsample height. The force signal from the Zwick machine was connected to the ARAMIS SRX, Zeiss (Jena, Germany) system controller. This made it possible to relate the tensile strain values to the force values. Figure 2 shows the measurement plan view from the software that operates the ARAMIS SRX system.
Measurements of geometry, displacement, and strain can be performed in the software. Strain measurements are performed using the DIC technique, which allows for the creation of measurement grids (Figure 3b) that are "tied" to the zone on which they are created. Correlation is possible by overlaying a small stochastic pattern of black and white points on the sample (Figure 3a) and dividing the whole into small subregions called facets.
The tests were carried out at a constant speed of 3 mm/min. Figure 4 shows the map of the principal deformation directions of the major strains (Eps1) and minor strains (Eps2) on the sample during bending. The Eps1 directions visualize where the material carries the largest strains and in which directions. It can be seen that the lower layers of the sample carry the strains horizontally in the X direction determined in the analyses. These layers are stretched. Similarly (Figure 4b) shows which layers are compressed. As in any pure bending test, there is a so-called neutral layer in the center of the cross-section that does not carry any strains.   The tests were carried out at a constant speed of 3 mm/min. Figure 4 shows the map 167 of the principal deformation directions of the major strains (Eps1) and minor strains (Eps2) 168 on the sample during bending. The Eps1 directions visualize where the material carries 169 the largest strains and in which directions. It can be seen that the lower layers of the sam-170 ple carry the strains horizontally in the X direction determined in the analyses. These 171 layers are stretched. Similarly (Figure 4b) shows which layers are compressed. As in any 172 pure bending test, there is a so-called neutral layer in the center of the cross-section that 173 does not carry any strains.  The tests were carried out at a constant speed of 3 mm/min. Figure 4 shows the map 167 of the principal deformation directions of the major strains (Eps1) and minor strains (Eps2) 168 on the sample during bending. The Eps1 directions visualize where the material carries 169 the largest strains and in which directions. It can be seen that the lower layers of the sam-170 ple carry the strains horizontally in the X direction determined in the analyses. These 171 layers are stretched. Similarly (Figure 4b) shows which layers are compressed. As in any 172 pure bending test, there is a so-called neutral layer in the center of the cross-section that 173 does not carry any strains.

180
The crosslinking reaction of the epoxy resin Epidian 601 using TETA resulted in rigid 181 polymeric composites. To fully crosslink the systems, the samples were kept at 70 °C for 182 1 h in a heating chamber. The reaction was carried out in special molds and then samples 183 were cut for mechanical testing. Six systems differing in the amount of natural filler (kao-184

Scanning Electron Microscope (SEM)
A Quanta 250 FEG (Field Emission Gun) scanning electron microscope (SEM) manufactured by FEI (Hillsboro, OR, USA) was used to investigate the morphology of the structure of the composite.

Results and Discussion
The crosslinking reaction of the epoxy resin Epidian 601 using TETA resulted in rigid polymeric composites. To fully crosslink the systems, the samples were kept at 70 • C for 1 h in a heating chamber. The reaction was carried out in special molds and then samples were cut for mechanical testing. Six systems differing in the amount of natural filler (kaolinite (KA) or clinoptilolite (CL)) were obtained. A sample without filler was also obtained and taken as the reference sample.
The presence of free hydroxyl groups on the surface of both fillers causes these materials to be characterized by high polarity, and therefore hydrogen bonds occur between the OH groups and the associated water molecules. The introduction of such a polar filler to an epoxy resin, whose epoxy group is also polar in nature, will result in an increase in the number of hydrogen bonds between the surface OH groups of the filler and the oxygen atoms of the epoxy group, as well as an increase in the ether bond of the resin. The occurrence of such hydrogen bonds is described in detail by Maity et al. [31]. It should be noted that such newly formed hydrogen bonds will not introduce new peaks in the ATR/FTIR spectra.
The above analysis indicates that the addition of a filler to the polymer matrix (pure EP601 resin) will not significantly affect the ATR/FTIR spectra of the obtained composites. It results from the fact that the main absorption bands of the fillers largely overlap with the bands of pure resin, which makes them invisible in the spectra of the composites.
The ATR/FTIR spectra of the materials obtained (see Figure 6 [4].
The FTIR spectra obtained after the DMA analysis did not show significant differences in the intensities of most bands, except for the band at about 915 cm −1 , the intensity of which decreased slightly. This suggests a reduction in the number of free epoxide groups as a result of the cross-linking of the samples during heating.

DSC Analysis
The changes the in the physical transformation of the materials obtained were determined by DSC analysis. The numerical data of the analyses are presented in Table 2, while the shapes of the DSC curves are presented in Figure 7. vibrations at about 3050 cm . In the range of 3050-3035 cm there are also asymmetric 238 C-H stretching vibrations of the resin epoxy ring, while deformation vibrations of the en-239 tire epoxy ring in the ATR/FTIR spectra are characterized by absorption bands at about 240 915 cm −1 [4]. 241 The FTIR spectra obtained after the DMA analysis did not show significant differ-242 ences in the intensities of most bands, except for the band at about 915 cm −1 , the intensity 243 of which decreased slightly. This suggests a reduction in the number of free epoxide 244 groups as a result of the cross-linking of the samples during heating.

248
The changes the in the physical transformation of the materials obtained were deter-249 mined by DSC analysis. The numerical data of the analyses are presented in Table 2, while 250 the shapes of the DSC curves are presented in Figure 7.   In the DSC curves from the first heating cycle (see Figure 7a), two energy effects are visible for all materials. The first one, in the range of 20-60 • C, is associated with the glassy transformation of the materials tested. The determined T g values are in the range of 32-54 • C. Analyzing the influence of the filler addition on the T g value, it is difficult to observe clear dependencies. In the series of materials based on kaolinite, the T g of the materials increases with the increase in the amount of filler, while in the series with clinoptilolite, there is no such relationship. The second energy effect visible in the DSC curves from the first heating, which is an exothermic effect, can be characterized as an effect related to the cross-linking of materials. This process took place at cross-linking temperatures (T cur ) between 107 and 117 • C, and the associated heat release (∆H) was in there is no such relationship. The second energy effect visible in the DSC curves from the 262 first heating, which is an exothermic effect, can be characterized as an effect related to the 263 cross-linking of materials. This process took place at cross-linking temperatures (Tcur) be-264 tween 107 and 117 °C, and the associated heat release (ΔH) was in the range of between 265 3.16 and 18.33 J/g. The cross-linking process of the obtained materials in both series did 266 not depend on the amount of filler in the sample. 267 The DSC curves from the second heating revealed only glass transitions (Tg in the 268 range of 40-65 °C) which were higher than those determined in the first heating. This is 269 consistent with the common knowledge that the cross-linking process shifts the glass-270 transition temperatures of the material towards higher temperatures. 271 As in the first heating cycle, in the series of materials with kaolinite, the Tg values 272 increased with the increase in the amount of filler. 273 Comparing the effect of the type of filler (kaolinite and clinoptilolite) on the parame-274 ters determined by the DSC method, it can be concluded that the type of filler does not 275 affect the thermal values of the composites obtained.  Table 3. The DSC curves from the second heating revealed only glass transitions (T g in the range of 40-65 • C) which were higher than those determined in the first heating. This is consistent with the common knowledge that the cross-linking process shifts the glasstransition temperatures of the material towards higher temperatures.
As in the first heating cycle, in the series of materials with kaolinite, the T g values increased with the increase in the amount of filler.
Comparing the effect of the type of filler (kaolinite and clinoptilolite) on the parameters determined by the DSC method, it can be concluded that the type of filler does not affect the thermal values of the composites obtained.

DMA Analysis
To investigate the influence of the fillers on the viscoelastic properties of composites, dynamic mechanical thermal analysis was performed. Changes in storage modulus (E ), loss modulus (E ), and mechanical loss factor (tan δ) with temperature are shown in Figure 8, while DMA data are summarized in Table 3. As can be seen from Table 3, the values of the storage modulus (E 20 ) characterizing the stiffness of the material increase with the increase in the filler content in the sample. Similar relationships can be observed for the loss modulus, which characterizes the elasticity of the tested materials. It is worth noting here that in the case of the loss modulus, the addition of 10% by weight of the filler increased the flexibility, while the other composites were characterized by lower E max values.   Table 3, the values of the storage modulus (E′20) characterizing 285 the stiffness of the material increase with the increase in the filler content in the sample. 286 Similar relationships can be observed for the loss modulus, which characterizes the elas-287 ticity of the tested materials. It is worth noting here that in the case of the loss modulus, 288 the addition of 10% by weight of the filler increased the flexibility, while the other com-289 posites were characterized by lower E″max values. 290 The flexibility of the material is also characterized by the value of the damping factor 291 (tan δmax). The higher its value, the less flexible the material is, which makes it less capable 292 of damping vibrations. Analyzing the effect of the filler addition on this parameter, it can 293 be seen that its addition increased the tan δmax value with respect to the starting material. 294 In the case of this parameter, there are clear differences in the type of filler used. In the 295 series of composites with kaolinite, the tan δmax values decreased as the filler content in-296 creased, while in the series with clinoptilolite this relationship was reversed. 297 The addition of the filler in both series increased the homogeneity (FWHM values) of 298 the composites obtained, compared with the material without the filler. This is interesting 299 because, in general, the addition of any component to the polymer mixture reduces its 300 homogeneity. In our case, the fillers used were to some extent bound to the polymer ma-301 trix through hydrogen bonds, as described in the section describing the ATR/FTIR spectra. 302 The flexibility of the material is also characterized by the value of the damping factor (tan δ max ). The higher its value, the less flexible the material is, which makes it less capable of damping vibrations. Analyzing the effect of the filler addition on this parameter, it can be seen that its addition increased the tan δ max value with respect to the starting material. In the case of this parameter, there are clear differences in the type of filler used. In the series of composites with kaolinite, the tan δ max values decreased as the filler content increased, while in the series with clinoptilolite this relationship was reversed.
The addition of the filler in both series increased the homogeneity (FWHM values) of the composites obtained, compared with the material without the filler. This is interesting because, in general, the addition of any component to the polymer mixture reduces its homogeneity. In our case, the fillers used were to some extent bound to the polymer matrix through hydrogen bonds, as described in the section describing the ATR/FTIR spectra.
Interesting conclusions confirming the results of the DSC analysis can be drawn after analyzing the temperatures of the glass transitions. Using the DMA method, we determined the T g values of materials from the maximum tan delta (tan δ max ) and maximum loss modulus (E max ). T g values from the loss modulus maximum were close to the T g values determined by the DSC method, while values from tan δ max were about 10-15 • C higher.
The shape of the storage modulus curves in Figure 8 correlates with the shape of the DSC curves from the first heating cycle. After the glass transition, which is shown by the inflection of the storage modulus curve, there is no flattening, but a rising of the line. This is due to the cross-linking of the sample during heating, which increases its stiffness. This effect is visible in both series of materials obtained.
However, when we trace the shape of the loss modulus and tan delta curves, two transformations can be observed. The first one (with maxima between −60 and −50 • C) is connected with the main chain movements of the polymeric matrix and local movements of polar groups. The second type of relaxation (called primary relaxation) with maxima between 25 and 70 • C is related to the glass transition of the polymer. The intensity of the principal relaxation peak determines the material's ability to dampen (absorb) vibrations; the smaller its value, the greater the material's damping capacity.
In Figure 9, the photos of the cut samples after DMA analysis are presented. As can be seen in the photo, heating the samples to 127 • C did not damage the samples.
However, when we trace the shape of the loss modulus and tan delta curves, two 314 transformations can be observed. The first one (with maxima between −60 and −50 °C) is 315 connected with the main chain movements of the polymeric matrix and local movements 316 of polar groups. The second type of relaxation (called primary relaxation) with maxima 317 between 25 and 70 °C is related to the glass transition of the polymer. The intensity of the 318 principal relaxation peak determines the material's ability to dampen (absorb) vibrations; 319 the smaller its value, the greater the material's damping capacity. 320 In Figure 9, the photos of the cut samples after DMA analysis are presented. As can 321 be seen in the photo, heating the samples to 127 °C did not damage the samples. 322 323 Figure 9. Samples after DMA analysis. 324

325
For the obtained materials, the hardness was tested before and after the DMA analy-326 sis, and the results are presented in Figure 10.

Mechanical Properties
For the obtained materials, the hardness was tested before and after the DMA analysis, and the results are presented in Figure 10. effect is visible in both series of materials obtained. 313 However, when we trace the shape of the loss modulus and tan delta curves, two 314 transformations can be observed. The first one (with maxima between −60 and −50 °C) is 315 connected with the main chain movements of the polymeric matrix and local movements 316 of polar groups. The second type of relaxation (called primary relaxation) with maxima 317 between 25 and 70 °C is related to the glass transition of the polymer. The intensity of the 318 principal relaxation peak determines the material's ability to dampen (absorb) vibrations; 319 the smaller its value, the greater the material's damping capacity. 320 In Figure 9, the photos of the cut samples after DMA analysis are presented. As can 321 be seen in the photo, heating the samples to 127 °C did not damage the samples. 322 323 Figure 9. Samples after DMA analysis. 324

325
For the obtained materials, the hardness was tested before and after the DMA analy-326 sis, and the results are presented in Figure 10. The materials obtained were characterized by Shore hardness values before DMA analysis in the range of 78.5-80.5 ShD, and after DMA analysis in the range of 80.0-81.5 ShD. As Figure 10 shows, the hardness of the materials increased with the increase in the content of the filler, and the type of filler did not affect the differences in hardness. In all cases, the DMA analysis increased the hardness of the samples. This is consistent with earlier observations from DSC and DMA analyses, which showed that after heating the samples, their cross-linking occurred.
The increase in hardness of the composites after DMA analysis may result from several factors, the occurrence of which was dependent on the heating of the samples. As already mentioned, when the samples were heated above 110 • C, additional chemical cross-linking took place, which resulted in the opening of the epoxide ring and the formation of new hydroxyl groups. Apart from cross-linking, the water molecules associated on the surface of the filler were also released during heating, which increased the number of free hydroxyl groups. This made the surface more polar. A greater number of hydroxyl groups on the surface of the filler and in the polymer matrix resulted in an increase in the amount of hydrogen bonds between the phases and thus in the hardness as well as stiffness of the obtained composites.
In Table 4 the bending and tensile strain data are presented. The DIC method also enables the observation of deformation maps, allowing for a visual determination of which zone experiences the highest deformation and where the material is under the most stress. Table 5 presents EpsX deformation maps at the stage just before sample destruction. One sample for each tested type was selected.
To compare the influence on the strength properties of the percentage of added KA and CL. the results were collated into two groups (Figures 11 and 12). Due to the insignificant difference in the cross-sectional area of all samples (the standard deviation in this range is 0.88 mm 2 compared to the mean cross-sectional area of all samples at the level of 35.71 mm 2 ), the analysis of bending force vs. deformation was chosen instead of bending stress vs. deformation.
In the case of kaolinite (Figure 11), the addition at the level of 5% and 10% slightly worsens both properties analyzed. However, in the case of the amount at the level of 1%, material plasticization occurs, which leads to the possibility of carrying deformations over 16 times greater than the reference material (25% deformation compared to 1.5% deformation). This is visually presented in Figure 13. material is under the most stress. Table 5 presents EpsX deformation maps at the stage 3 just before sample destruction. One sample for each tested type was selected.
3 material is under the most stress. Table 5 presents EpsX deformation maps at the stage just before sample destruction. One sample for each tested type was selected. material is under the most stress. Table 5 presents EpsX deformation maps at the stage just before sample destruction. One sample for each tested type was selected. material is under the most stress. Table 5 presents EpsX deformation maps at the stage just before sample destruction. One sample for each tested type was selected. just before sample destruction. One sample for each tested type was selected.  To compare the influence on the strength properties of the percentage of added KA and CL. the results were collated into two groups (Figure 11 and Figure 12). Due to the insignificant difference in the cross-sectional area of all samples (the standard deviation in this range is 0.88 mm 2 compared to the mean cross-sectional area of all samples at the level of 35.71 mm 2 ), the analysis of bending force vs. deformation was chosen instead of bending stress vs. deformation. To compare the influence on the strength properties of the percentage of added KA and CL. the results were collated into two groups ( Figure 11 and Figure 12). Due to the insignificant difference in the cross-sectional area of all samples (the standard deviation in this range is 0.88 mm 2 compared to the mean cross-sectional area of all samples at the level of 35.71 mm 2 ), the analysis of bending force vs. deformation was chosen instead of bending stress vs. deformation.
To compare the influence on the strength properties of the percentage of added KA 355 and CL. the results were collated into two groups ( Figure 11 and Figure 12). Due to the 356 insignificant difference in the cross-sectional area of all samples (the standard deviation 357 in this range is 0.88 mm 2 compared to the mean cross-sectional area of all samples at the 358 level of 35.71 mm 2 ), the analysis of bending force vs. deformation was chosen instead of 359 bending stress vs. deformation.  In the case of kaolinite (Figure 11), the addition at the level of 5% and 10% slightly 365 worsens both properties analyzed. However, in the case of the amount at the level of 1%, 366 material plasticization occurs, which leads to the possibility of carrying deformations over 367 16 times greater than the reference material (25% deformation compared to 1.5% defor-368 mation). This is visually presented in Figure 13. Level 0% in Figure 13 was taken as the reference material value. A positive percentage change indicates an improvement in properties, while a negative change indicates a deterioration in properties.

SEM Analysis
In Figure 14 the SEM images of the composites being studied before and after mechanical tests are presented.  In the case of kaolinite (Figure 11), the addition at the level of 5% and 10% slightly 365 worsens both properties analyzed. However, in the case of the amount at the level of 1%, 366 material plasticization occurs, which leads to the possibility of carrying deformations over 367 16 times greater than the reference material (25% deformation compared to 1.5% defor-368 mation). This is visually presented in Figure 13. Level 0% in Figure 13 was taken as the reference material value. A positive percent-374 age change indicates an improvement in properties, while a negative change indicates a 375 deterioration in properties. 376

377
In Figure 14 the SEM images of the composites being studied before and after me-378 chanical tests are presented. Studies of the microstructure of the composites obtained using scanning electron mi-383 croscopy (SEM) show the different nature of the two mineral additives. The surfaces of 384 the reference sample are characterized by a homogeneous internal structure (Figure 14a). 385 The addition of both kaolinite and clinoptilolite causes the distinct appearance of thin 386 overlapping layers in the microstructure of the materials tested. The structure of these 387 composites becomes less homogeneous, and pear-shaped aggregates of the applied addi-388 tives are visible (Figure 14b and 14c). This is particularly evident in composites whose 389 composition was modified by the addition of kaolinite. Analyzing the surfaces obtained 390 as a result of destructive forces in composites containing in their composition the addition 391 of kaolinite or clinoptilolite, it is clear that the process of destruction took place on the 392 mineral aggregates present in the structure of the tested materials ( Figure 14b' and 14c'). 393

394
During the crosslinking reaction, the new composites with advantageous thermal 395 and mechanical properties are obtained from the epoxy resin (Epidian 601) with triethyl-396 enetetramine and inorganic mineral fillers. The FTIR analysis confirms the opening of the 397 epoxy ring reaction, as the decrease of 915cm −1 signal in the composite spectra is visible. 398 DSC analysis shows that the type of filler (kaolinite and clinoptilolite) does not affect the 399 thermal values of the composites obtained. The Tg values of the materials obtained are in 400 the range of 30-54 °C, which is typical for crosslinked epoxy resins. 401 The addition of the filler to polymer matrices caused an increase in the homogeneity 402 as well as stiffness, as indicated by dynamical thermal analysis. As can be seen from DSC 403 and DMA analysis, upon heating the materials obtained became harder. It indicates the 404 crosslinking effect on the samples after increasing the temperature. 405 By analyzing the effect of the percentage addition of clinoptilolite on strength prop-406 erties, it can be unequivocally stated that for the added value of 10% the properties dete-407 riorated. Both the maximum force carried by the sample in the bending test and the max-408 imum tensile strains in the lower layer of the sample are lower than those of the reference 409 material. However, at addition values of 1% and 5% there is a slight improvement in the 410 maximum force value and a slight decrease in the strain values for 1% clinoptilolite, and 411 strain values for 5% clinoptilolite are similar to those in the reference material. Studies of the microstructure of the composites obtained using scanning electron microscopy (SEM) show the different nature of the two mineral additives. The surfaces of the reference sample are characterized by a homogeneous internal structure (Figure 14a). The addition of both kaolinite and clinoptilolite causes the distinct appearance of thin overlapping layers in the microstructure of the materials tested. The structure of these composites becomes less homogeneous, and pear-shaped aggregates of the applied additives are visible (Figure 14b,c). This is particularly evident in composites whose composition was modified by the addition of kaolinite. Analyzing the surfaces obtained as a result of destructive forces in composites containing in their composition the addition of kaolinite or clinoptilolite, it is clear that the process of destruction took place on the mineral aggregates present in the structure of the tested materials (Figure 14b',c').

Conclusions
During the crosslinking reaction, the new composites with advantageous thermal and mechanical properties are obtained from the epoxy resin (Epidian 601) with triethylenetetramine and inorganic mineral fillers. The FTIR analysis confirms the opening of the epoxy ring reaction, as the decrease of 915 cm −1 signal in the composite spectra is visible. DSC analysis shows that the type of filler (kaolinite and clinoptilolite) does not affect the thermal values of the composites obtained. The T g values of the materials obtained are in the range of 30-54 • C, which is typical for crosslinked epoxy resins.
The addition of the filler to polymer matrices caused an increase in the homogeneity as well as stiffness, as indicated by dynamical thermal analysis. As can be seen from DSC and DMA analysis, upon heating the materials obtained became harder. It indicates the crosslinking effect on the samples after increasing the temperature.
By analyzing the effect of the percentage addition of clinoptilolite on strength properties, it can be unequivocally stated that for the added value of 10% the properties deteriorated. Both the maximum force carried by the sample in the bending test and the maximum tensile strains in the lower layer of the sample are lower than those of the reference material. However, at addition values of 1% and 5% there is a slight improvement in the maximum force value and a slight decrease in the strain values for 1% clinoptilolite, and strain values for 5% clinoptilolite are similar to those in the reference material.
The confirmed appropriate thermal, thermo-mechanical, and mechanical properties of the materials obtained indicate the possibility of their application f.ex. in the building industry.