INVESTIGATION INTO THE INFLUENCE OF FLAME RETARDANT ADDITIVES ON SOME FIRE PROPERTIES OF POLYESTER MATERIALS APPLYING SMALL-SCALE TESTING TECHNIQUES

In order to investigate the impact of some inorganic additive flame retardants on the selected fire properties of the materials based on polyester resin Polimal 1033 APy, small-scale fire testing techniques have been used. Seven samples have been studied: unmodified PES, PES modified with MoO3 (7, 14 and 21 wt%) and PES modified with Sb2O3 (7, 14 and 21 wt%). The following flammable properties of materials have been determined: the heat of combustion (HOC), the ignition temperature of volatile thermal decomposition products (Tig), self-ignition temperature and oxygen index. A cone calorimeter method has been used for determining heat release rate (HRR), mass loss, specific extinction area (SEA) and other combustion parameters. The toxicological analysis of combustion products has been conducted. Based on the obtained results, the following conclusions have been made: (1) MoO3 and Sb2O3 added to the studied material change its flammable properties and fire parameters. It can be indicated by higher HOC, higher Tig and self-ignition temperature, as well as by lower HRR and SEA. Modified materials become safer in terms of fire hazard. (2) A significant reduction in HRRmax of approx. 40% in the content of 7 wt% has been observed. The lowest HRRmax of approximately 300 kW/m 2 and 450 kW/m have been obtained for 21 wt% in a range of 200 600 s at 30 kW/m and 100 400 s at 50 kW/m respectively. Except for a sample containing 7 wt% of Sb2O3, a clear local reduction in HRR (from 50 to 150 kW/m ), in case of all modified samples has been noticed. (3) Sb2O3 has a greater impact on the thermostability of the studied materials compared to MoO3 in all cases of heat flux density and additive concentrations. The effectiveness of Sb2O3, as a flame retardant is the most evident at 21 wt%.


Introduction
Polyester materials are of high significance among many types of materials applied nowadays in building engineering (building elements, division walls, glass roofs, structural elements, floor screeds). They are also used in textile industry, automotive industry, electronics, etc.
Traditional methods for reducing flammability are based on the modification of plastics by adding flame retardants, the so called antipyrenes, for example, as curing agents or fillers (Gałaj et al. 2012;Pofit Szczepań ska, Pó łka 2002a, b, 2003Pó łka 2001). Flame retardants inhibit the process of the thermal decomposition of the material subjected to thermal radiation resulting in a reduction in material combustion rate. A consequence of such process, among the others, is lower material flammability and a slower increase in fire temperature and CO concentration creating more favorable conditions for evacuation from the building. Basically, two mechanisms of acting antipyrenes, including physical and chemical, can be distinguished (Jankowska et al. 2007;Wilkie, Morgan 2009). The physical effect of the flame retardant agent mainly consists of the gas dilution phase and the formation of barrier layers that block the mass flow of thermal decomposition products and energy flow between gas and solid phases during material combustion. The chemical effect consists of antipyrene entering into free radical reactions of the combustion process in the gas phase and/or into reactions occurring in the solid phase.
A disadvantage of some flame retardants is frequently an increase in the toxicity degree of thermal decomposition products and the combustion of modified polymers in the fire environment. The products formed during combustion also pose danger to people during evacuation. Thus, due to the toxicity of the products of thermal decomposition and combustion, in recent years, the agents consisting of halogen atoms have been often substituted by halogen-free inhibitors.
A large number of papers present the analyzed results of different mechanisms for fire retardancy of polymer materials and small-scale fire tests using standard methods, such as a cone calorimeter, thermogravimetric analysis (TGA), limited oxygen index (LOI), a differential scanning calorimeter (DSC) and/ or a vertical burning test (UL 94) (Carosio et al. 2012;Fukushima et al. 2010;Kandare et al. 2008;Kiliaris, Papaspyrides 2010;Konecki, Pó łka 2006, 2009aLaoutid et al. 2009;Li et al. 2012;Pan et al. 2012;Pereira et al. 2009;Pó łka 2001;Seetapan et al. 2011;Tang et al. 2012;Tibiletti et al. 2011;Wei et al. 2011). The main directions of research in the considered field of the study include the determination of susceptibility to initiate combustion reactions, particularly under thermal and flow conditions simulating the 1 st phase of fire, the quantitative and qualitative composition of thermal decomposition products reflecting smokegenerating capacity and the toxicity of the fire environment, the formation rate of critical heat flux during the combustion of epoxides enabling the transition of the 1st fire phase into the 2 nd active phase, as well as the influence of the physical and chemical parameters of additives on the flammability of polyesters.
In order to investigate the behavior of polyester material based on Polimal 1033 Apy, the experimental part of the study on thermal and flammable properties looks at unmodified and modified inorganic additives during fire occurrence, as well as reveals heat release rates at the selected heat flux according to ISO 5660Á 1:2002(Konecki, Pó łka 2005Półka 2001). The toxicity of materials according to PNÁ88/BÁ02855 (1988) has been also investigated.

Subject of the study
For experimental small-scale studies by means of the cone calorimeter method, a trade name of polyester resin (PES) Polimal 1033 APy produced in the Chemical Plant Organika Sarzyna in Nowa Sarzyna, Poland and commonly applied in building engineering has been used. Structural resin has low viscosity and is medium elastic, orthophtalic and pre-accelerated. An abbreviation at the end of the name stands for ecological, low styrene emission (letter A) and accelerated resin and does not require the use of cobalt accelerator (symbol Py). The basic parameters of resin are provided in Pó łka (2001).
For modifying Polimal 1033 Apy resin, two inorganic oxides were used: antimony trioxide (Sb 2 O 3 ) and molybdenum trioxide (MoO 3 ). The size of the mesh Sb 2 O 3 (a) was up to 23 mm; however, MoO 3 made of 2Á10 mm was produced by POCH S. A. Gliwice (Chemical Corporation in Poland). Modifications are aimed at enhancing fire resistance properties of resin. The quantities of added antipyrenes have been selected experimentally. Concentration range of additives depended on visible changes in a particular fire property of the lowest additive concentration and a lack of registered change in the same parameter of fire above the highest applied concentration of the additive. For selecting inorganic filler, a change in resin viscosity and gelation time was taken into account. The method of modification for both additives was the same. Both additives were added to polyester resin obtained as a result of polycondensation at a temperature of t025 8C and pressure p00 atm (Pó łka 2001).
For small-scale studies, pure PES polyester and polyesters modified adding flame retardants reducing their flammability were used. The content of the additives in the modified samples was 7 wt%, 14 wt% and 21 wt%. Pure PES was transparent. The material with molybdenum addition was green, while a sample with the addition of antimony was white.

Test stand and measuring methods
The thermal analysis of unmodified and modified polyester materials was conducted in dynamic conditions. Thus, the mass of the studied decomposed sample was registered as a function of temperature at a constant heating rate. For thermogravimetric studies, the shredded samples of polyester material (of approx. 0.5 mm) made of unmodified Polimal 1033 Apy and modified with inorganic oxides having the mass of 140Á170 mg were used. Thermal analysis was performed applying a derivatograph manufactured by Hungarian company MOM Á Budapest (Magyar Optikai Muvek). The samples were placed in a thermoanalyzer oven and tested in the temperature range of 20Á1000 8C at three different heating rates ( 2.5, 5 and 10 8C/min. Analyses were conducted under atmospheric conditions. In order to investigate the flammable properties of seven analyzed polyester materials, the following parameters were determined: combustion heat, the ignition temperature of volatile products of thermal decomposition, self-ignition temperature and oxygen indexes of the studied polyesters. The heats of combustion analysis were carried out based on PNÁ 81/GÁ04523 (1981). The chosen method consisted of the complete combustion of the weighed sample of a solid material in oxygen atmosphere under pressure in a bomb calorimeter and the measurement of temperature growth in water in the calorimetric vessel. For studying combustion heat, the samples of polyester products were used Á shredded to the fraction of 1 mm of 1 g mass. The ignition temperature of the volatile products of thermal decomposition was determined according to PNÁ69/CÁ89022 (1969). The method consists of heating the sample up to temperature lower than 10 8C of the expected thermal decomposition temperature and testing the ignition (during 5 min) of emitted pyrolytic gases with reference to an auxiliary ignition source (most frequently flame). Polyester material samples shredded to the fraction of 0.5 mm (mass 01 g) were placed in a metal thimble and put into an oven under constant temperature in the range of 150Á400 8C. Self-ignition temperatures of unmodified samples and samples modified with inorganic oxides were determined based on the method described in ASTM D 1929Á77 (1977. The method consists of placing the sample in the oven with adjustable temperature and measuring the lowest temperature of the oven walls at which the selfignition of thermal decomposition products occur. This self-ignition temperature can be determined under both static and dynamic conditions. The dynamic method was applied (Stechkin apparatus CSÁ88 by Custom Scientific Instruments) in the presented studies more accurately reflecting fire conditions. The studies focused on the air conditioned samples of polyester material having the dimensions of 20)20 mm and the mass of 390.5 g. The determination of oxygen indexes was conducted based on PNÁ76/CÁ89056 (1976) standard. The method of oxygen index consists of determining the lowest oxygen concentration in the oxygen and nitrogen mixture, at which a plastic sample fixed vertically in the measuring column burns for 3 minutes at a length of 5 cm. The samples of polyester materials with a shape of rectangular beams and dimensions of 80)1) 4 mm were tested employing standard test equipment manufactured by Stanton Redcroft.
For analyzing heat release rate and other combustion parameters, a cone calorimeter manufactured by Fire Testing Technology (Great Britain) was applied. Its general view is given in Fig. 1a. The tests were conducted according to ISO 5660Á1:2002 standard. The samples of the studied polyester materials were subjected to the influence of the heat flux of 30 kW/m 2 and 50 kW/m 2 . To select the power of heat radiation to determine fire parameters, a thermal flux simulating the 1st and 2nd phase of fire was used. Ignition was initiated by piloted ignition. The ignition of volatile products was performed by means of a spark igniter. The studies employed a horizontal placement of the tested samples with regards to the radiator (see Fig. 1b). Before testing, unmodified and modified polyester material samples having the dimensions of 100)100 mm were stored at 2092 8C, wrapped in aluminum foil and placed in the test frame for cone calorimeter studies. A surface of the samples subjected to heat flux impact was 88.4 cm 2 , the thickness of the profiles was 4Á5 mm, and the initial weight of the samples varied by a small percentage.

Studies on the results of thermal properties
The results of the thermal properties of polyester material samples are given in Table 1. The thermal distribution of the studied materials is a function of their composition and heating rate. Unmodified polyester in the range of the studied temperature and at applied heating rates does not change its thermostability. The temperatures of starting thermal decomposition and maximum mass loss, influencing mainly the ignition temperatures of the studied materials vary slightly, which indicates that the variable heating rate of polyester material based on Polimal 1033 APy at the beginning of the 1 st phase of fire, does not influence the rate of forming the fire environment. Modification with flame retardants such as molybdenum or antimony oxides has an impact on the thermostability of the material. The results of thermogravimetric studies shown in Table 1 indicate an inhibitive impact of both applied oxides, while significant effect has heating rate along with additive concentration. Taking into account three applied concentrations of additive Á molybdenum trioxide (7, 14 and 21 wt%), a maximum increase in the decomposition temperatures of 30 8C was observed in the case of sample PES '21% MoO 3 compared to unmodified polyester. For the maximum mass loss of the studied modified PES, the high concentration of the additive has a significant influence. At molybdenum trioxide content of 21 wt% found in PES, maximum mass loss decreases from 75 wt% (for unmodified PES) to 46 wt%, which significantly reduces the concentration of the flammable volatile phase and consequently reduces the risk of fire hazard posed by clear polyester material. In case of applying Sb 2 O 3 as a modifier of PES flammability, changes in thermostability are more apparent. The thermostability of tested polyesters increases at all applied concentrations of antimony trioxide and heating rates. The highest growth in both temperatures is that the beginning of thermal decomposition and maximum mass loss can be observed at Sb 2 O 3 concentration of 14 wt% and the heating rate of 10 8C/min.
The maximum mass loss of the material based on Polimal 1033 APy has been demonstrated at Sb 2 O 3 content of 21 wt% (59%). It should be noticed a significant variation of growth in the temperature of maximum mass loss at applied heating rates. An addition of antipyrenes increases the amount of resulting residues after pyrolysis. The highest amount of ashes after sample combustion can be observed in case of polyester material containing 21 wt% of MoO 3 and the heating rate of 2.5 8C/min. A comparative analysis of the influence of both tested additives on the studied polyester material, based on the obtained thermogravimetric results, demonstrated antimony trioxide to be a more effective modifier of the flammable properties of the products, based on Polimal 1033 APy rather than molybdenum trioxide increasing its thermostability. Table 2 presents ignition temperatures, the heat of combustion, as well as self-ignition and oxygen indexes obtained during testing the samples of polyester material. The flammable properties of the studied polyester material both modified and unmodified with flame retardants, were analyzed based on the obtained ignition temperatures of thermal decomposi- tion products, self-ignition temperatures, oxygen indexes and the heat of combustion. Three first values indicating the ability to initiate a combustion reaction listed above were determined applying standard methods using low-energy heat sources. The heat of combustion was determined employing the standard method of material combustion in oxygen. The obtained results of ignition temperature and self-ignition indicate that the impact of both additives on the flammability of the volatile phase, formed as a result of the thermal decomposition of PES, is insignificant taking into account the initiation of combustion that uses the MoO 3 modifier. The maximum difference in the temperature values of ignition regarding unmodified and modified polyester material was 7 8C at 21 wt% content of molybdenum oxide in PES and 12 8C at 21 wt% content of antimony trioxide. In case of self-ignition temperatures of the studied materials, the maximum difference in the values of 38 8C has been observed for PES including 21% of Sb 2 O 3 and 20 8C for PES with 21% of MoO 3 . Most probably, in case of MoO 3 , it is related to the low volatility of antipyrene, which, as demonstrated in further results, inhibits especially in the condensed phase slightly changing the composition of the volatile phase of the unmodified polyester material. Therefore, the values of both the temperature of the ignition of the volatile products of thermal decomposition and self-ignition temperature, before and after modification are similar. Greater effects of the differentiation of susceptibility for ignition have been demonstrated using Sb 2 O 3 , although antimony trioxide transits into the volatile state much easier, as a result of the thermal decomposition of polyester material, comparing to molybdenum trioxide. The obtained results indicate that the composition of the volatile phase formed during the thermal decomposition of the material modified with Sb 2 O 3 has a lower ability of flame propagation either in the case of point heat source or heat radiation. The above observations have been also confirmed by the results of measuring the oxygen index (IO) of polyesters. However, mass and heat transfer are different during oxygen index measurements comparing to mass and heat transfer between heat source and material during studies on ignition and self-ignition temperatures. IO values vary insignificantly and thus the studied polyester both unmodified and modified can be still classified as easy to ignite (IO B21%). Combustion heat values of the studied material with inorganic fillers are lower in comparison with the values for unmodified materials. Along with an increase in the concentrations of the applied oxides, combustion heat values are decreasing. The lowest value has been observed for resin containing 21 wt% of Sb 2 O 3 . As the effect of adding molybdenum and antimony trioxide and as the antipyrenes of polyester materials, the material made of Polimal 1033 Apy becomes safer in terms of fire hazard and reflects lower fire loading of the buildings in which polyester materials are used as building materials. The above feature allows applying the materials of a lower fire resistance in buildings.

Results and analysis of fire parameters and toxicity
One of the most important fire parameters and one of the main factors determining the rates of temperature growth and release of toxic products is heat release rate (HRR). The time when HRR reaches the maximum value is crucial; whereas the shorter the time, the greater is the risk for humans and the environment.
With reference to the conducted studies and applied software, the following fire characteristics and flammable properties of the material have been determined: (1) Heat release rate in kW/m 2 : Maximum heat release rate HRR max in kW/m 2 ; Time for reaching HRR max Á T HRRmax in s; Mean heat release rate HRR mean in kW/m 2 ; (2) Mass loss rate MLR in g /(m 2 s): Mean mass loss rate MLR mean in g (m 2 s); (3) Total heat released THR in MJ/m 2 ; (4) Effective heat of combustion HOC in MJ/kg: Mean effective heat of combustion HOC mean in MJ/kg; (5) Specific extinction area SEA in m 2 /kg: Mean specific extinction area SEA mean in m 2 /kg; (6) Final sample mass FSM in g; (7) Time for constant ignition T ig in s. The results of conducted studies in the form of selected thermo-physical and thermo-kinetic parameters of studied PES for two different external types of heat flux density Á 30 and 50 kW/m 2 are shown in Tables 3Á6. Figs 2Á5 present HRR curves for pure and modified PES with pilot ignition at varying different heat flux density of 30 and 50kW/m 2 .
Based on the conducted cone calorimeter studies, the following observations have been made: (1) The addition of 7 wt% of MoO 3 and Sb 2 O 3 to polyester material Polimal 1033 Apy reduces heat release rate for approx. 40Á 50% at the analyzed heat flux density (see Tables 3 and 5 and Figs 2Á5); (2) Along with an increase in filler concentration, HRR max and HRR mean of materials under conditions of ignition are decreasing thus reaching the lowest values of 21 wt% (see Tables 3 and 5); (3) Anitmony trioxide more effectively reduces heat release rate from polyester material than molybdenum trioxide. HRR max values of the products containing Sb 2 O 3 are lower for about 20% compared to the values of the products modified with MoO 3 under conditions of ignition and at the studied heat flux density (see Tables 3 and 5); (4) Independently from MoO 3 concentration and heat flux density, HRR curves of PES 'MoO 3 have two peaks and a specific 'saddle', which indicates the formation of a carbonized layer on the surface of the material. The carbonized layer is a barrier for incoming oxygen and heat, and prevents the spread of the combustion process. The greatest difference in height between two peaks has been observed for PES '7 wt% of MoO 3 . HRR curves of polyester materials containing Sb 2 O 3 are characterized by more stable HRR values compared to PES with MoO 3 in the whole combustion range. The above indicates a lack of the carbonized layer or the formation of a very small layer (see Figs 2Á5); (5) Along with an increase in the concentration of both flame retardants, heat release rate from modified PES is decreasing (see Tables  3 and 5 Tables 3 and 5); (7) Maximum heat release rates for unmodified and modified PES for applied heat flux density depend on the method of initiating the combustion process. The combustion theory distinguishes three types of combustion initiation: ignition, self-ignition and spontaneous combustion. Due to the fact that unmodified and modified polyester material does not undergo spontaneous combustion, the studies used two other types determining fire occurrence. Under conditions of selfignition, HRR max values of the studied materials are lower comparing to the corresponding materials under conditions of the ignition of pyrolytic products. The difference in HRR max between materials arising from various types of the combustion initiation process is 14% at 30 kW/m 2 and 15Á20% at 50 kW/m 2 ; under conditions of self-ignition and heat flux density of 30 kW/m 2 , time for sample self-ignition is longer for 17% than time for the self-ignition of the unmodified material. Time for the ignition and self-ignition of the studied materials depends not only on the type and concentrations of fillers, but also on the heat release rate and type of combustion process initiation. At a lower heat flux of 30 kW/m 2 , time for the ignition or selfignition of the studied PESs is longer than at the flux of 50 kW/m 2 (see Tables 4 and 6 Tables 4 and 6); (11) The mean mass loss rates (MLR mean ) of the studied polyester materials under ignition and self-ignition conditions decrease along with an increase in the concentrations of the applied fillers. The highest value of MLR mean has been observed for unmodified PES. The thermal decomposition of crosslinked Polimal 1033 Apy including Sb 2 O 3 proceeds faster than with MoO 3 ; it is indicated by higher values of MLR mean for materials containing Sb 2 O 3 in the studied heat fluxes independently from the combustion initiation method. Considering the initial and final mass of the studied samples, it should be emphasized that, in most cases, residual mass after pyrolysis is larger for materials containing MoO 3 . Investigated materials containing Sb 2 O 3 under the influence of incident heat flux mostly demonstrate lower (comparing to the one containing MoO 3 ) ash content after combustion (see Tables 3 and 5); (12) Flame retardants Sb 2 O 3 and MoO 3 added to polyester resin increase the mean coefficient of the specific extinction area (SEA) of the material. The studied polyesters containing MoO 3 at the given heat flux and under ignition conditions demonstrate a lower SEA value of approx. 15% compared to the materials containing Sb 2 O 3 (see Tables 3 and 5).
Studies on polyester resins have indicated that the additives do not have a significant influence on changes in toxicity. Toxicological profiles of concentrations for the studied substances with regards to temperature allows noticing that along with temperature growth CO concentration drops and CO 2 increases. It has been shown that adding the antimony trioxide of 7 wt% results in a slight increase in toxicity, whereas in case of the sample consisting of 14 wt%, toxicity drops reaching the best result from all studied materials.
The main conclusion resulting from the analysis of the obtained results is that all studied materials were classified as moderately toxic thus being at the lowest safest limit according to PNÁ88/BÁ02855 (1988). The received results are presented in Tables 7Á12.
Definitions of toxic indicators: Á toxic indicator W LC50 [g/m 3 ] Á the weight of the given material, the decomposition or burning of which under testing conditions generated toxic threshold concentration of the given decomposition or burning product: where: LC 50 [g/m 3 ] Á threshold concentration of the given thermal decomposition product that causes the death of 50% of the population at an exposition of 30 minutes; E Á specific emission [g/g] (mean arithmetical value of the specific emission value in view of the results of at least two tests); Chemical decomposition and burning products according to PNÁ88/BÁ02855 (1988) are classified taking into account appropriate groups presented in Table 13.

Summary and general conclusions
The paper focuses on the results of small-scale studies considering the evaluation of the flammability of new polyester materials based on Polimal 1033 APy resin, in which special flame retardants, such as molybdenum and antimony trioxides found in different content (7 wt%, 14 wt% and 21 wt%) are used. The obtained results have been compared with data received from pure resin without any additives.
With reference to the analysis of the obtained results, the following conclusions have been formulated.
Molybdenum and antimony trioxides introduced to polyester resin change the flammable properties and fire characteristic of the studied polyester material. Together with the increasing content of inorganic oxides, the modified materials based on Polimal 1033 APy become safer from the point of view of fire hazard. It is indicated by higher values of heat combustion, lower heat release rates of the modified materials under considered heat flux and in consequence of a lower fire load of the buildings, where structural elements are made of non-saturated polyester resins. Higher values of the ignition temperature of volatile products and self-ignition and time for ignition or self-ignition confirm the above conclusion, since polyester materials containing flame retardants MoO 3 and Sb 2 O 3 inhibit the initiation of the combustion reaction, and thus inhibit fire development at its initial phase. An inhibitive impact of antimony trioxide on self-ignition and ignition temperatures and time for the self-ignition and ignition of PES is greater comparing the above temperatures and time for the material containing molybdenum trioxide. The effectiveness of Sb 2 O 3 as antipyrene for polyester materials is the most evident when Sb 2 O 3 concentration is 21 wt%. Antimony trioxide has a greater influence on the thermostability of the investigated material based on Polimal 1033 APy compared to molybdenum trioxide in all cases of heat flux density and all additive concentrations. There is a significant diversification of temperature growth in maximum mass loss in light of the studied heating rates, which indicates that the thermal decomposition of the studied polyester materials is a function of their composition and heating rate. The temperatures of the initial thermal decomposition of polyesters modified with oxides are higher compared to unmodified materials. The addition of inorganic fillers to PES  results in a growth in the temperatures of maximum mass loss during thermal decomposition and simultaneously in the increased amount of residues after pyrolysis. The mass of ashes obtained from polyester samples containing MoO 3 in view of the studied heat flux density is higher compared to unmodified materials and those containing Sb 2 O 3 in considered concentrations. The inhibitory effect of Sb 2 O 3 on the studied material is related to both the condensed phase and the volatile phase. It is demonstrated by longer time for the ignition and self-ignition of the material consisting of Polimal 1033 Apy containing Sb 2 O 3 , compared to the material composed of MoO 3 clear PES at the heat flux density of 30 and 50kW/m 2 . Antimony trioxide included in PES during the heating process shows greater tendency to transit into the volatile phase than molybdenum trioxide and indicates inhibiting time for ignition. Molybdenum trioxide applied in PES only slightly influences changes in time for the ignition or self-ignition of modified materials under studied conditions. Toxicity studies resulting in the determination of toxicometric indicators have demonstrated that both unmodified and modified PES are moderately toxic. Therefore, the influence of additives on the toxicity of the studied materials is insignificant. However, it should be emphasized that the lowest toxicity among all introduced samples (W LC50SM 0139.6 g/m 3 ) was observed in PES containing 14 wt% of antimony trioxide.
The applied flame retardants even at the concentration of 7 wt% cause a significant reduction in heat release rate for approx. 40% (from 700 kW/m 2 to 400 kW/ 2 ). A further increase in the content of additives, though to a lesser degree of approx. 30%, further decreases HRR values. The lowest HRR max values of about 300 kW/m 2 at the heat flux of 30 kW/m 2 and 450 kW/m 2 and the radiation source of 50 kW/m 2 were obtained in the concentration of 21 wt%.
In the range of 200Á600 s at 30 kW/m 2 and 100Á 400 s at 50 kW/m 2 , with the exception of the sample containing 7 wt% of Sb 2 O 3 , a visible local reduction in heat release rate in case of all modified samples (from 50 to 150 kW/m 2 ) can be observed and indicates the formation of the carbonated zone on the surface, which is a barrier to incoming oxygen and heat, thus reducing flame propagation.
During combustion initiated by spark ignition in case of the sample with no additives, heat release rate constantly increases very rapidly in the first phase and slows down up to the maximum value within next 200 s. In case of the samples containing additives, after approximately 200 s for 30 kW/m 2 and after 100 s for 50 kW/m 2 and following a rapid increase in HRR, the value stabilizes approximately at the level of 250 kW/m 2 at the heat flux of 30 kW/m 2 or at the level of 300Á400 kW/m 2 at the heat flux of 50 kW/m 2 , and, in most cases, forms a specific 'saddle'. Moderately toxic