Investigation of the thermal protection ablative properties of thermosetting composites with powder fillers: the corundum Al2O3 and the Carbon Powder C

Abstract Phenol formaldehyde resins were filled with mixtures of corundum Al2O3 and carbon C powders to produce thermoprotective composites. The composites were treated with hot combustion gases to determinate the temperature profiles across rectangular samples of dimensions 10x25x35 mm. The carbonization of the thermosetting matrix was observed. It was qualified the qualitative and the quantitative effect of components on the ablation surface temperature, the back side temperature of specimen and the mass waste under intensive heat fl ow after 120 s of treatment with hot combustion gases. The composites with higher matrix content (more liquid resin and less adhesive resin) and with predominance of corundum Al2O3 over carbon powder C showed the best insulating properties.


INTRODUCTION
Modifi ed plastics were introduced as ablative protecting materials against excessive temperature in the middle of the 20 th century, fi nding applications in the arms industry as well as aeronautical, rocket and space industries 1 . These materials can also be used in the design of passive fi re--proof protections for large cubature supporting elements in building structures 2 , communication tunnels 3-4 and for the protection of data stored in electronic, optical and magnetic carriers 5 .
Autonomous shields and ablative shielding are also used to protect building structures and people's lives in heat load incidents with temperatures much exceeding permissible standards. The classical fl ameproof materials cannot prevent the increase in temperature at the rear side of protective shields as effectively as ablative materials whose high characteristics of substitute heat resistance r kp allows the reduction of temperatures from several dozen degrees to ~ 2 000°C with the use of relatively thin insulation shields 1, 5- 9 .
Endothermic reactions associated with thermal decomposition of polymer matrix take place in polymer composites at temperatures higher than the ablation temperature t a , which lead to high values of effective specifi c heat c p . In their pure form, polymers are considered to be good ablative materials 1, 5-9 having very low heat conductivity coeffi cient λ. However, due to their softening, low density and low heat stability as well as brittleness which are induced during the ablation of ablative layer, there is a need to combine these materials with mineral powders or fi ber reinforcements which improves the thermal stability, resistance to heat fl ux and consequently the thermal insulation of composites.
Despite many years of experience with ablative materials, the relationship between the phases type and composition with ablative properties, within the context of others operational properties of the composites used as thermal protection shields, remains still not evaluated qualitatively and quantitatively 5, 6, 9 .
The present work investigates the effect of quantitative and qualitative phase composition of phenol formaldehyde composites with chosen powder fi llers (corundum Al 2 O 3 and carbon C) on ablative thermo-protective properties of these composites, especially on the back side temperature of specimen t s and the mass waste U a .

Materials
Materials were chosen on the basis of the thermo--physical properties of powder fi llers, their commonness in applications to provide thermal protection as well as good ablative properties, high thermal stability, incombustibility and lack of toxic products of thermal decomposition of the resin. Based on bibliography 5-14 the following materials have been used to prepare the specimens of polymer composites: thermosetting matrix, resins (Modofen 54S and Nowolak MR); powder fi llers, corundum Al 2 O 3 (ALO G5-4) with grains of 2 to 5 μm with the minimal contents of aluminium oxide of 99.5% (95% α Al 2 O 3 ) manufactured by Ajka Amumina, Hungary and fi ne grain carbon powder C of 5 μm and purity of 98% manufactured in Poland.

Samples preparation
Experimental samples for thermoprotective investigations were created in a six-step procedure: 1. Preparing a mixture of fi llers (homogenization of the mixture of both powders).
2. Adding appropriate mass of grain damper (liquid resin Modofen 54S increasing adhesion of the binder to fi llers) and mixing it to homogenize the mixture.
3. Adding weighted amount of the binder powder (powder resin Nowolak MR) and mixing everything until uniform, powdery material is formed. One obtains a semi-fi nished product consisting of loose groups of grains covered with two resin coatings: grains of the fi llers in the inner coating of liquid resin (the grain damper) are covered with loose powder (the binder).
4. Filling the metal grid with the obtained powder and increasing its density under pressure of approximately 0.4 MPa.
5. Closing sides of the metal mould and keeping it in 150°C for 60 minutes to allow gelation of the resin, fi ller supersaturation and crosslinking the matrix. Moulds had air channels allowing for degasifi cation of the hardening samples.
6. Cooling and removing of the forms to obtain an experimental sample.
The components of the response variable y (the output parameters) are the average maximal back side temperature of specimen t s (max) [ o C] and the average mass waste U a [%] after 120 s of treatment with hot combustion gases.
The regression coeffi cients b i of all function components have been calculated. The statistical analysis of the tests results allowed for the determination of the threshold relevance of the regression coeffi cients b i and estimation of their effect on the output parameters y. The output value is signed in the equation of the experiment objective (1) 15 : Moreover, the variance , error in determination regression coeffi cients s(b i ) and their level of statistical signifi cance b istot have been determined on the basis of t-Student test at confi dence level 95% 15 .

Evaluation of ablative properties
The so-called "ablating gun" 16 (Fig. 1) of own construction 5, 7 was used for the classical tests of ablative thermal protection properties, enabling thus the interaction of steady and uniform streams of infl ammable gases on the samples at high temperatures. The specimens (of size 10x25x35 mm) were placed in a shielding made of fl ameproof plaster-cardboard panel and exposed to gas heat fl ux for τ = 120 seconds. Burning of Methyl Acetylene and Propadiene mixture (MAPP gas) in an ablating gun blowpipe provided the source of heat (Fig. 2).
The registration of decomposition temperatures on the ablative surface t pa (Fig. 3 -typical sample) and back surface t s (Fig. 4 - Czaki Thermo-Product). Furthermore, the ablation weight loss of composites U a has been also evaluated. The temperature on the ablative surface increased more rapidly the fi rst 10 s, followed by a low rate of temperature increase for the rest of the test (Fig. 3). Fast increase of temperature during fi rst 10 seconds is caused by rapid heating of the sample surface. Next, temperature grows slower, because apart from taking the heat and conducting it deeper into the sample, endothermic ablation processes start (heat is used i.e. to decompose resins).
The temperature of the back side increased with increasing time of ablation. Specimens 7 and 8 exhibited the highest and lowest back side temperature (Fig. 4).
The average results of ablation tests and the relation of both response variables (the output parameters) to the phase variables are presented in Table 2 and their graphic interpretations in Figure 5.
Shaping the thermal protective ablative properties of polymer composites can be done by choosing such phase composition which: guarantees the possibly most signifi cant reduction of the temperature on the thickness of the protective wall; ensures the emergence of the ablative layer which is coherent, thermally stable, characterized by good adhesion to the basic material with a low coeffi cient of thermal conductivity; increases the thermal stability of the basic material and its endurance to withstand thermal-mechanical strain which appears because of the impact of the heat fl ux; ensures the lack of combustibility, smoke and toxicity of the appearing ablation products.
The aim of the experiment was to fi nd such a composite whose values of the average maximal back side temperature t s (max) and the average mass waste U a are the lowest and it gives good reduction of temperature and good thermal stability of the material. These conditions have been met by specimen 8 whose phase composition consists of 40% matrix (28% Modofen 54S and 12% Nowolak MR), 48% corundum Al 2 O 3 , and 12% carbon powder C (Fig. 5).

STATISTICAL ANALYSIS OF TEST RESULTS
The regression coeffi cients and their signifi cance, the variance , and the determination error s(b i ) were calculated based on the data provided in Table 2 and were presented in Table 3. The bold print marks b i -values, which is lower than b istot but burdened with the error s(b i ) which allows to assess the b i as statistically signifi cant. The lack of data in the Table means that the given index is Table 2. The factual phase compositions of the composites and the results of ablation tests  Having analysed the values and preceding signs of the regression and interaction coeffi cients, we can confi rm that for the assumed range of independent variables, there exist the relation of the thermoprotective parameters to the coding variables that is to the phase composition of the composite: With the increase in matrix contents, the temperature t s (max) decreases, but the weight loss U a increases (coeffi cient b 1 ).
The increase in the corundum share Al 2 O 3 (the decrease in the carbon powder share C) results in the decrease of ablative weight loss U a , but at the same time leads to the increase in temperature t s (max) (coeffi cient b 3 ).
When the increase in the matrix content is related to the increase in the Modofen 54S share (negative value of b 12 in case of t s (max) ), with the higher content of Al 2 O 3 and smaller content of carbon powder C (negative value of b 23 in the case of U a as well as negative values of b 13 in the case of t s (max) and U a ), this type of phase composition of the composite ensures the top thermo-protective properties, that is the lowest temperature t s (max) and the lowest ablative weight loss U a.

STRUCTURES OF ABLATIVE SAMPLES
Experimental samples after the ablation process were covered with chemically crosslinked resin in cylindrical moulds (plastic tubes). After hardening the resin, it was cut along the heat fl ux direction in the symmetry axis of the ablative surface. Samples were cut mechanically at slow speed and with cooling to avoid temperature increase in the cross-section, which could affect the structure and thermal stability of the polymer. Next, the samples were wet polished using sandpaper with decreasing grain size and then wet polished using felt with polishing paste (corundum). No corrosives were used. Then pictures of the ablative layer structures and the basic material were taken using an optical microscope.
In Figure 6, the border between the basic and secondary ablation is visible (mark by an arrow). In the structure of the ablative layer we can see the so-called vitreous slagthe porous substance of low thermal conductivity, which is the solid product of pyrolysis and ablation processes (the dark area of the secondary ablation). It is visible below the congealed/solidifi ed ceramic compounds (the light area of the secondary ablation). Above it, one can observe a congealed, fi ne-grain ceramic structure which is the transition as well as secondary reactions zone, almost free from solid products of pyrolysis. It can be stated that the resin matrix of the composite got decomposed below the border of the secondary ablation. Figure 7a shows the composite basic material, which has not become the subject of ablation. One can easily see big, light, angular grains of corundum Al 2 O 3 against the background of black carbon powder and structural defects -pores which appeared in the technological process.
In Figure 7b, we can see the fi ne-grained structure of solidifi ed ceramic (secondary ablation) with a typical dendrite type of grain (light areas), against the dark background of porous, glassy slag (the substance that remained after the secondary ablation). Instead of having sharp edges, the grains of ceramic have ovoid ones, which proves that they emerged in result of chemical reactions from substrates of the secondary ablation process (on the basis of carbon and corundum) and then got solidifi ed due to the termination of the impact of the heat fl ux.

CONCLUSIONS
The composites containing higher amount of matrix (more liquid resin Modofen 54S and less adhesive resin Nowolak MR) have the best thermo-protective ablation proprieties when corundum Al 2 O 3 is predominant in the fi ller mixture (over carbon powder C). Both temperature of specimen t s and average mass loss U a are the lowest (tested sample 8).
The kind and the content of powder fi llers are the most important phase composition parameters, as they  give the highest effect on the improvement of thermal stability and resistance to heat fl ux. The higher amount of corundum Al 2 O 3 (the lower content of carbon powder C) the lover ablation weight loss of composites U a . The statistical analysis -for the fi rst order experimental plan -allows us to quantify the infl uence of each level of the factor variables on both response variables. The signs of the regression coeffi cients show if the response is increasing or decreasing, while their absolute values inform us about the strength of the dependence. Coupling this information with technological constraints infl uencing the values of the input variables indicates the way towards optimal phase composition of ablative thermal-protective composites.
The ablation layer is porous. In the zone of secondary reactions it has fi ne-grained dendrite structure made of ceramic compounds, which were created from fi llers and products of polymer pyrolysis of the composite matrix under high-temperature heat fl ux. In the primary ablation zone we observe high-melting grains of fi llers covered in solid products of resin pyrolysis.