Influence of Fire Retardants and Nanofillers on Fire Toxicity

CHAPTER

Influence of Fire Retardants and Nanofillers on Fire Toxicity

24 A.A. Stec, T.R. Hull

Centre for Fire and Hazard Science, University of Central Lancashire, Preston, UK

CHAPTER OUTLINE 1. Introduction ....................................................................................................... 837 1.1 Fire statistics ...................................................................................... 839 1.2 Fire scenarios and combustion conditions ............................................. 839 1.3 Experimental methodology ................................................................... 841 1.4 Toxic potency of fire effluent ................................................................ 842 1.5 Quantification of environmental pollution from fire effluents, and hazards from particulates and their impact on human health............................... 845 2. Analysis of fire toxicity for different polymers and products .................................. 847 2.1 Toxic product yields and estimated fire toxicity of common polymers ....... 847 2.2 Effect of fire retardants and nanofillers on fire toxicity ............................ 850 2.2.1. Effect of fire retardants and nanoclays on fire toxicity of polypropylene .............................................................................. 850 2.2.2. Effect of fire retardants on ethylene-vinyl acetate copolymer ............. 850 2.2.3. Effect of fire retardants on polyester materials .................................. 851 2.2.4. Effect of fire retardants and NCs on PAs .......................................... 853 2.3 Quantification of effluent toxicity during burning of fire retarded commercial products (precise formulation unknown) .............................. 856 2.3.1. Medium density fiber board ............................................................. 856 2.3.2. Insulation products .......................................................................... 857 2.3.3. PVC carpet ...................................................................................... 860 2.4 Cables and cable materials .................................................................. 862 3. Conclusions ....................................................................................................... 863 References ............................................................................................................. 864

1. Introduction Most fire deaths and most fire injuries result from the inhalation of toxic smoke [1], both of which have increased with the widespread use of plastic materials. Fire toxicity is most important in areas where escape is restricted. Thus, most mass transport applications, such as airlines, railways, and passenger ships include requirements Polymer Green Flame Retardants. http://dx.doi.org/10.1016/B978-0-444-53808-6.00024-X Copyright © 2014 Elsevier B.V. All rights reserved.

837

838

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

Fire hazard assessment

Materials

Fire characteristics Ventilation

Enclosure

Temperature

Toxic product yield

Asphyxiants

Irritants

CO, HCN, CO

HCl, HBr, NO, NO

Smoke

Geometry Heat

Available safe escape time

FIGURE 1 Schematic of factors required for fire hazard assessment relating to fire toxicity.

to quantify the fire toxicity of internal components. Fire hazard assessment requires consideration of the most probable fire scenarios, prediction of the rate of fire growth, the amount of fuel present, and its impact on the occupants and their ability to escape safely [2]. Figure 1 shows a schematic relationship between the factors required to assess the fire hazard, in order to ensure that the time required for escape is less than the time before escape becomes impossible (available safe escape time) [3]. The burning of an organic material produces a cocktail of products that vary with fire conditions. In addition to carbon dioxide (CO2), water (H2O), and oxygen (O2) depletion these include asphyxiant gases, carbon monoxide (CO), and hydrogen cyanide (HCN); irritant gases, hydrogen chloride (HCl), hydrogen bromide (HBr), nitrogen oxides (NO, NO2), and organoirritants, such as acrolein and formaldehyde, and particulates. They can cause death directly by asphyxiation or indirectly by inhibiting breathing (gaseous irritants and particulate irritants) or by visual obscuration (smoke), in each case preventing escape. Prediction of toxic fire hazard depends on two parameters: 1. Timeeconcentration profiles for major products. These depend on the fire growth curve and the yields of toxic products. 2. Toxic potency of the products, based on estimates of doses likely to impair escape efficiency, cause incapacitation, or death. While some real-life fires may be represented by a single fire stage, most fires progress through several different stages [4]. Burning behavior and particularly toxic product yields depend most strongly on a few factors, especially material composition, temperature, and oxygen concentration [5]. The generalized development of a

1. Introduction

fire has been recognized, and used to classify fire growth into a number of stages, from smoldering combustion and early well-ventilated flaming, through to fully developed underventilated flaming [6].

1.1 Fire statistics Over the last 50 years, there has been a continuous change in the materials used for construction and content of buildings and transport with an increasing dependence on synthetic polymers and their composites with different reaction-to-fire properties. Compared with natural materials (wood, wool, cotton, leather, etc.), widely used synthetic polymers such as polyethylene (PE), polypropylene (PP), polyamide (PA), polystyrene (PS), and polyurethane (PU) burn more quickly, and generate more smoke and toxic effluents particularly in low-density, open structures such as fabrics and foams. Although the overall number of deaths has decreased [1] as smoke alarms and fire detection systems became available, UK Fire Statistics, with the best time series data in the world, show a progressive shift in the cause of death from “burns” to “overcome by toxic gas or smoke” from 1955 to 2008. More remarkable is the very large rise in fire toxicity injuries [1]. These are shown in Chapter 4 in Figures 1 and 2. In Europe, most victims are found in the room of fire origin, often close to an exit, where they collapsed trying to escape. Usually, these are living rooms or bedrooms, with upholstery or bedding being the first thing to catch fire. EU directives for zero energy homes (e.g. the Passive House level (2012e2015) and the Zero Carbon level (2016e2020)) may result in millions of homes being lined in flammable insulation materials such as PS and PU foam (PUR) or equally flammable “environmental” materials such as flax or shredded paper, and this may have a dramatic impact on domestic fire deaths.

1.2 Fire scenarios and combustion conditions Toxic product yields depend on the interactions between the material composition and the fire conditions; particularly temperature and ventilation and whether decomposition is nonflaming or flaming [7]. As the fire develops, the conditions change: the temperature increases and oxygen concentration decreases. The temperature and oxygen concentration vary significantly during a fire, and between different fires. Different fire scenarios are shown in Figure 2. The most important fire stages identified by the ISO, from nonflaming to wellventilated flaming to underventilated flaming, have been classified in terms of heat flux, temperature, oxygen availability, and CO2 to CO ratio, equivalence ratio 4 and combustion efficiency (the % conversion of fuel to fully oxygenated products, such as CO2 and water). The most significant differences in terms of toxic product yields with fire conditions arise between flaming and nonflaming combustion. Typically, nonflaming combustion produces small volumes of highly toxic, partially burnt effluent, well-ventilated fires produce larger volumes of less toxic effluent, while developed underventilated fires

839

840

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

FIGURE 2 Stages of fire growth.

produce large volumes of toxic effluent, and therefore represent the greatest hazard. For flaming combustion, the fuel/air ratio has the greatest effect on the yields. Research predicting the carbon monoxide evolution from flames of simple hydrocarbons, reviewed by Pitts [8], has shown the importance of the equivalence ratio, f, for predicting the CO yield from the oxygen depletion in flaming conditions, presented in Table 1. Stoichiometric combustion describes burning a material in its nominal chemical oxygen requirement (f ¼ 1) providing just enough oxygen for full oxidation to CO2 and H2O. In practice, for condensed phase fuels, very little air usually results in incomplete combustion because mixing is never perfect. In well-ventilated fires, combustion is fuel lean with excess air (f between 0.5 and 0.75), while in the later stages of a fire, when there is not enough air available and the conditions are underventilated, the equivalence ratio will be typically between 1.5 and 2.5. Each fire stage has a characteristic temperature and equivalence ratio, as summarized in Table 2.

Table 1 Equivalence Ratio f¼

actual fuel to air ratio stoichiometric fuel to air ratio

f¼1 f1 f>1

“Stoichiometric” combustion Well-ventilated fires (fuel lean flames) Underventilated fires (fuel-rich flames)

Table 2 Classification of Fire Stages ISO Fire Stage

Temperature/ C

Equivalence Ratio 4

1a. Oxidative pyrolysis 2. Well-ventilated flaming 3. Under ventilated flaming: 3a. Small fires 3b. Post-flashover fires

350 650

Not applicable f < 0.75

1–5 2–20

650 825

f > 1.5 f > 1.5

2–20 2–20

CO2/CO Ratio

1. Introduction

Fire can also be characterized by the CO2/CO ratio as an indicator of combustion efficiency. However, this only correlates with the equivalence ratio (and degree of ventilation) when there is no gas phase quenching e.g. by hydrogen halides. Therefore, this parameter also corresponds to combustion efficiency.

1.3 Experimental methodology Ideally, all bench-scale fire effluent toxicity test methods should be capable of reproducing the conditions in each of the stages of actual fires, including incipient, growing, and fully developed fires. It is therefore essential for the assessment of toxic hazard from fire that each fire stage can be adequately replicated, and preferably the individual fire stages treated separately [2,9]. Several test methods have been used to generate products for the purpose of evaluating the toxic product yields from burning polymers. However, many of them fail to relate this to a particular fire scenario [2,10,11]. In addition, although room- and larger-scale fire tests have been also conducted and the results published [4,12e14], only a few of these have attempted to segregate the fire stages, allowing the complexities of full-scale burning behavior to be addressed using a bench-scale model. The steady-state tube furnace, ISO 19700 [15,16], has been developed specifically to replicate individual fire stages by decomposing materials under the full range of fire conditions from oxidative pyrolysis to fully developed, underventilated flaming [17,18]. By controlling the fuel feed rate and air flow into a tubular furnace, steady burning can be achieved for different fuel/air ratios even forcing combustion in oxygen-depleted atmospheres [19]. The apparatus is shown in Figure 3. The apparatus typically consists of a tube furnace (600- to 800-mm heating zone) and a quartz tube (1600e1700 mm) that passes through the furnace and into a mixing and measurement chamber (27e30-l capacity). The standard procedure uses around 20 g of material uniformly placed in a silica boat (800 mm) to give a

FTIR Secondary air supply

Sample boat driven in over ~20 min

Primary air supply

F nace Fur Furnace ac ce e

FIGURE 3 The steady-state tube furnace, ISO 19700.

Smoke Smo oke mea measurement asurem men nt

Exhaust gases

Secondary oxidiser (for hydrocarbons)

841

842

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

linear density of 25 mg mm1. A drive mechanism pushes the specimen boat into the furnace tube at a rate of 40 mm min1 to give a fuel introduction rate of 1.00 g min1. A constant stream of primary air is provided at the furnace tube entry, and secondary air is supplied into the mixing chamber to give a total air flow through the apparatus of 50 l min1. This provides a steady-state mass charge concentration of 20 mg l1 (or 20 g m3) in the chamber. After dilution with secondary air, this effluent is equivalent to a fuel mass of 1 kg in 50 m3 in a room. For materials that leave a residue or form a char, the residual mass is measured, and product yields based on mass loss concentrations may be also calculated. The requirement in each test run is to obtain a steady state of at least 5 min during which the concentrations of effluent gases and particulates can be measured. A light/photo cell system is used to determine smoke density across the mixing and measurement chamber. Organics (unburnt and partially burnt hydrocarbons) are determined as products of incomplete combustion using a secondary oxidizer for further oxidation at 900  C in excess air, over silica wool, as the difference between secondary CO2 and primary CO and CO2 measured using a nondispersive infrared analyzer. The toxicity of organic species in the fire effluent may be quantified as a ratio of the actual organic yield to the organic yield of 10 mg l resulting in incapacitation, as described by Purser [4]. All scale-up of fire is difficult, but particularly in combustion toxicity where product yields may differ by two orders of magnitude, depending on fire scenarios [14]. It was shown that for many common materials, the yield of toxic products such as carbon monoxide, hydrogen cyanide, organoirritants, and smoke increases by a factor between 10 and 50 as the fire changes from well ventilated (0.5 < f < 0.7) to underventilated (1< f < 5). The steady-state tube furnace has been shown to replicate a range of large-scale fire stages or conditions, characterizing the fire behavior of materials under controlled and well-defined laboratory conditions, in terms of the equivalence ratio (f) or the CO2/CO ratio [20]. Further, since each test run represents the burning behavior for a particular fire stage, the results are better defined than those of a single large-scale test, where individual fire stages may coexist so the transition is indistinguishable [21e23]. An example for CO from PP is presented in Figure 4. The large-scale test data (from the ISO room corner test) show a good agreement with the steady-state tube furnace data; a significant improvement on any other toxicity test. In well-ventilated conditions, both the tube furnace and the largescale fire give CO yields of around 0.02e0.03 g g1, then following the same rising trend to 0.1 g g1 at f ¼ 1.3 in the large-scale test and 0.17 g g1 at f ¼ 1.5 in the tube furnace and for underventilated flaming.

1.4 Toxic potency of fire effluent Death or incapacitation may be predicted by quantifying the fire effluents in different fire conditions in small-scale tests, using chemical analysis. Lethality may be predicted using equations, based on rat lethality data, presented in ISO

1. Introduction

0.2

CO yield g g−1

Large scale Steady state tube furnace

0.15

0.1

0.05

0 0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Equivalence ratio ϕ

FIGURE 4 Comparison of tube furnace CO yields with large scale data for polypropylene [2].

13344 [24]. Incapacitation (the inability to effect one’s own escape) may be predicted using methodology and consensus estimate data in ISO 13571 [25]. Toxic product data from chemical analysis may be expressed in various ways, including effluent gas concentrations, effluent gas yields, toxicity indices, Fractional Effective Dose (FED), Fractional Effective Concentration (FEC), and LC50 (lethal concentration to 50% of the population). The general approach in generating toxic potency data from chemical analysis is to assume additive behavior of individual toxicants, and to express the concentration of each as its fraction of the lethal concentration for 50% of the population for a 30-min exposure (LC50) [4,24]. Thus, an FED ¼ 1 indicates that the sum of the actual/lethality concentration ratios of individual species will be lethal to 50% of the population over a 30-in exposure. Since CO2 increases the respiration rate, the Purser model, presented in Eqn (1), uses a multiplication factor for CO2 driven by hyperventilation, VCO2 , to increase the FED contribution from all the toxic species, and incorporates an acidosis factor A to account for toxicity of CO2 in its own right [24]. 
½CO ½HCN ½HCl ½NO2  þ þ þ þ . þ organics FED ¼ LC50;CO LC50;HCN LC50;HCl LC50;NO2 VCO2 þ A þ VCO2 ¼ 1 þ

21  ½O2  ; 21  5:4

expð0:14½CO2 Þ  1 ; 2

where A is an acidosis factor equal to½CO2   0:05: Equation (1) is the Purser model for estimation of toxicity of fire effluents.

(1)

843

844

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

This equation is related only to lethality, or “cause of death”. However, many people fail to escape from fires because of the incapacitating effect of smoke (obscuring visibility) and its irritant components that cause pain, inhibiting breathing, actually being the “reason for death”. ISO 13571 considers the four major hazards from fire that may prevent escape (toxic gases, irritant gases, heat, and smoke obscuration). It includes a separate calculation for prediction of incapacitation by each of the four hazards for humans exposed to fire effluents (indicating, in a nonnormative appendix that the effects of heat, smoke, and toxicants may be estimated independently). Equations (2) and (3) have been taken from ISO 13571. They calculate the FED of asphyxiants, CO and HCN, and the FEC of sensory irritants in the fire effluent that limit escape. t2 t2 X X ½CO expð½HCN=43Þ FED ¼ Dt þ Dt: (2) 35; 000 220 t1 t1 Equation (2) is the FED model for ISO 13571. Equation (2) considers the two significant asphyxiant fire gases CO and HCN. The FED value is calculated using the exposed dose relationship (concentrationetime product, C∙t) for CO. The lethal C∙t product corresponds to the incapacitating dose (C∙t) for CO of 35 000 ml l-1 min, equal to around 1170 ppm for a 30-min exposure) and an exponential relationship for HCN. FEC ¼

½HCl ½HBr ½HF ½SO2  ½NO2  ½acrolein þ þ þ þ þ IC50;HCl IC50;HBr IC50;HF IC50;SO2 IC50;NO2 IC50;acrolein ½fomaldehyde X ½irritant þ þ : IC50;fomaldehyde IC50;irritant

(3)

Equation (3) is the FEC model from ISO 13571. Equation (3) uses the same additive of Purser’s principle in ISO 13344 to estimate the combined effect of all irritant gases (IC is the concentration resulting in incapacitation of 50% of the population). ISO 13571 is a more robust methodology for ensuring the safety of potential fire victims. It is included here for completeness, to show the steps needed to avoid incapacitation, but subsequent results are presented using the simpler FED calculation of ISO 13344 (rat lethality) in order to illustrate the relative and absolute contributions of individual toxicants. In order to relate the fire effluent toxicity to a maximum permissible loading, the FED can be related to the mass of material in a unit volume, which would cause 50% lethality or incapacitation for a given fire condition; as an LC50, a specimen mass M of a burning polymeric material that would yield an FED equal to one within a volume of 1 m3 according to Eqn (4) [24]. LC50 ¼

M : FED  V

(4)

1. Introduction

Equation (4) is the relation of LC50 to FED.Here V is the total volume of diluted fire effluent in cubic meters at standard temperature and pressure (STP). Comparing the toxic potencies of different materials, the lower the LC50 (the smaller the amount of materials necessary to reach the toxic potency) the greater its fire toxicity. LC50 values should be referenced to the fire condition under which they were measured.

1.5 Quantification of environmental pollution from fire effluents, and hazards from particulates and their impact on human health Unwanted fires are characterized by incomplete combustion, often occurring at lower temperatures, and producing a rich cocktail of toxic and pollutant compounds compared to the small well-ventilated flaming and complete combustion of many laboratory test scenarios. In particular, real fires will produce more persistent bioaccumulative and toxic (PBT) products, such as polycyclic aromatic hydrocarbons (PAHs), and polychloro- and polybromo-dibenzo-dioxins and furans (PCDD/F and PBDD/F) from fuels containing halogens. In addition, research concerning exposure to respirable particles (from PM10s that are <10 mm diameter down to nanoparticles, w100 nm) from man-made sources indicates that exposure to background levels of particulates can cause irritation or damage to the respiratory system [26]. The World Health Organization estimates a 0.5% increase in daily mortality per 10 mg m3 of PM10 (particulate matter smaller than about 10 mm) and smaller particles [27]. The general effect of particulates is to cause fluid release and inflammation in the lungs, preventing gas exchange. The toxicity and particulate formation mechanisms present during accidental fires are not well understood, but the toxic gaseous species present in the smoke are known to attach themselves to nanoparticles, which act as vehicles taking toxicants deep into the lung. The particle size distribution is dependent on the material, temperature, and fire conditions. The particle size of the spherical droplets from smoldering combustion is generally of the order of 1 mm, while the size of the irregular soot particulates from flaming combustion is often larger, but much harder to quantify and dependent on the measuring technique and sampling position [20]. The deposition areas for humans as a function of particle size are shown in Figure 5. The PAHs produced from fires agglomerate together into small spherical particles, which then adhere to one another, like a tangled string of beads. Both the volatile PAHs and the spherical particles will remain airborne almost indefinitely, but have the potential to cause significant damage to the lungs. The lowmolecular-weight PAHs show a moderate potential to be adsorbed onto particulates whereas high molecular weight PAHs have stronger tendencies to adsorb to organic carbon, especially small particles (<2.5 mm) [28]. Of particular concern are the decomposition products from halogenated materials, including PVC and brominated flame retardants, which have been shown to produce large

845

846

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

FIGURE 5 Particulate distribution in the respiratory tract.

quantities of PCDD/F or PBDD/F in unwanted fires. PCDDs and PBDDs are extremely toxic, chemically and thermally stable, and have a tendency to be strongly adsorbed on the surface of particulate matter. The rate of formation of PCDD/Fs is a function of temperature, and the quantity of unburnt carbon, which is dictated by the oxygen level [29]. Numerous studies on animals have also confirmed that some dioxin congeners are carcinogenic and produce mutagenic effects in certain species. However, it is difficult to extrapolate data from different fire test methods as there are significant differences in both the combustion conditions (such as large variations in temperature and oxygen environments), and also varying concentrations of precursors and chlorine in the product gases [30]. Therefore, at present, our knowledge of the distribution and quantities of these species in fire gases is fairly limited. General awareness of the fact that fires may present dramatic and persistent adverse effects on the environment has been accentuated by a number of high impact incidents over the past half century [31]. Since quantitative data on environmentally hazardous components of fire effluent cannot be obtained from accidental fires, appropriate data must come from real-scale fire tests and simulations involving small-scale physical fire models. Such data are almost entirely absent from literature reports or enquiries relating to these incidents. Similar to PAHs, depending on the temperatures and oxygen regimes, the presence of unburned carbon particles in the postcombustion zone significantly increases the yield of PCDD/Fs.

2. Analysis of fire toxicity for different polymers and products

2. Analysis of fire toxicity for different polymers and products Estimation of fire toxicity is generally limited to those products considered most significant in causing incapacitation and death in fire victims. These consist predominantly of asphyxiant gases (CO, HCN, CO2, and lowered oxygen) and irritants, which include acid gases (HBr, HCl, NO2). In addition, examples of yield of carbon as organic carbon and carbonaceous particulates and other specific classes of compounds are also reported.

2.1 Toxic product yields and estimated fire toxicity of common polymers The yields of toxicants from polymethylmethacrylate (PMMA), low-density polyethylene (LDPE), PVC, polyamide 6.6 (PA 6.6), and PS are presented in Figure 6 as a function of equivalence ratio, f [2,32,33]. For PMMA, very efficient combustion at equivalence ratios below 1 is observed, with high CO2 yields and very low yields of other carbon-containing products. In underventilated combustion conditions, there is a large increase in yields of organic gases and CO with increasing f. For underventilated combustion, the yields of organics and CO are both high compared to well-ventilated flaming. A similar pattern of product yields for LDPE is observed. The theoretical yield (100% conversion of fuel carbon to CO2) for LDPE is 3.14 g/g under well-ventilated conditions. It can be seen that almost all the fuel carbon is converted to CO2. The CO2 yield follows an almost linear decrease as the ventilation is decreased, while the CO yield increases with decrease in ventilation from f ¼ 0.5 to f ¼ 1.5 then starts to decrease slowly, presumably due to limited oxygen availability. At an equivalence ratio of 1.6, the results show some increases in yields of hydrocarbons. PS has a similar carbon content to that of LDPE, but a higher carbon/hydrogen ratio. The combustion product yield pattern is similar to that of LDPE, but with less sensitivity to the ventilation conditions. It shows a high CO2 yield at low equivalence ratios, decreasing as f > 1, but there is a greater propensity to form carbon-rich soot throughout the f range and especially at high equivalence ratios [34]. The relatively high CO yields in well-ventilated combustion conditions and low CO yields in underventilated conditions suggest the presence of stable aromatic molecules, and low hydrogen ratio, resulting in inefficient oxidation. The higher CO yield and correspondingly higher soot yield for aromatics and unsaturated fuels burning in wellventilated conditions is well known; the lower CO yield under fuel-rich conditions is more interesting; the CO yield of 0.11 g g1 for toluene under fuel-rich conditions, and attributed it to the thermal stability of the molecule [35]. This results in a further reduction in combustion efficiency as the aromatic hydrocarbons are not converted to CO. Since the main product of decomposition of PS is the monomer, with smaller quantities of dimer, trimer, and tetramer, these are also likely to

847

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

1.5

1.0

0.2

0.0 0.5

1

1.5

2

0.5 CO, organics yields /g g−1

2.0

2.0 1.5 0.2

0.0

0.0

3.0

0.4

2.0

0.3

1.5

0.2

1.0

0.1

0.5

CO yields /g g−1

CO, organics yields /g g−1

0.5 0.0 0.5

1

1.5

Φ

2

2.5

0.6

2.5

CO, organics, HCN, NO yields /g g−1

CO organics as CH CO

0.5

1.0

0

Φ

PS

2.5

0.3

0.1

0.6

3.0

0.4

0.5

2.5

3.5

CO organics as CH CO

LDPE

0.4

0

0.6

2.5

CO yields /g g−1

CO, organics yields /g g−1

CO total organics as CH CO

3.0

Polyamide 6.6

CO organics as CH HCN NO NO CO

0.5 0.4

2.5 2.0

0.3

1.5

0.2

1.0

0.1

0.5

0

0.0

0.0 0

0.5

1

1.5

2

2.5

0.0 0

0.5

1

1.5

2

2.5

Φ

Φ 0.6

CO yields /g g−1

3.0

PMMA

CO yields /g g−1

0.6

3.0 CO organics as CH CO HCl

0.5

2.5

0.4

2.0

0.3

1.5

0.2

1.0

0.1

0.5

0

CO , HCl yields g g−1

PVC CO, organics yields g g−1

848

0.0 0

0.5

1

1.5

2

2.5

Φ

FIGURE 6 Fire effluent yields from polymethylmethacrylate (PMMA), low density polyethylene (LDPE), polystyrene (PS), polyamide (PA) 6.6, and polyvinyl chloride (PVC).

show similar enhancements in thermal stability, limiting the availability of OH$ radicals, responsible for converting CO to CO2. PA 6.6 shows a similar trend to PP and PMMA. Combustion is efficient at low equivalence ratios with low yields of CO and organics, all of which increase at equivalence ratios >1. The efficiency of conversion of fuel carbon to CO is lower than that for PMMA, which may be related to the lower oxygen content of the polymer. The yields for CO and HCN show an increase with increase in f. In wellventilated combustion conditions, the yield was low, but increased steeply to a maximum of 0.44 g g1 at f ¼ 2.5. Under well-ventilated combustion conditions, the main toxic species produced was NO at a maximum yield of 0.012 g g1 at

2. Analysis of fire toxicity for different polymers and products

f ¼ 0.5, decreasing as f increased to a yield of 0.0026 g g1 in underventilated combustion conditions. The yield of NO2 was approximately a factor of 10 lower than that of NO [10,17]. PVC burns with a low heat release rate, because the halogen atoms in the structure are released as HCl, accounting for almost 60% of its mass, which then inhibits the conversion of CO to CO2 and combustion is inefficient, even under wellventilated conditions [36]. The yield pattern is very different from that of all the other polymers described, in that the yields of all products are relatively similar across the whole f range from 0.5 to 2.5. Combustion is very inefficient across the range, with relatively low CO2 yields and high yields of CO, and organics. The CO yields in the steady-state tube furnace, are almost constant with increase in f. The theoretical yield of HCl is 0.585 g g1, and it can be seen that most of the chlorine is released as HCl. The major LDPE, PMMA, PS, PA 6.6, and PVC product yields, obtained from the steady-state tube furnace, have been translated into an overall estimation of the fire effluent toxicity, using the methods described in ISO 13344 indicating the contribution of each toxicant towards the overall fire hazard. Hypoxia is presented here as a decrease in oxygen supplied to, or utilized by, body tissue. The toxicity is expressed as FED for a fuel mass charge concentration of 20 g m3. Most polymers without heteroatoms follow the trend shown by LDPE and PMMA of fire toxicity increasing from a very low value in well-ventilated conditions, to a much higher value in underventilated flaming. As can be seen from Figure 7, there is a large variation in FED values for materials containing chlorine or nitrogen. For well-ventilated tests, the largest FED value is for PVC. The FED 10.0

CO

8.0

HCN

Hypoxia

NO2

HCl

Organic

FED

6.0

4.0

2.0

PMMA

LDPE

FIGURE 7 Fractional effective dose (FED) for common polymers [2].

PVC

PA 6.6

Large-UV

Well-V

Small-UV

Oxidative pyrolysis

Large-UV

Well-V

Small-UV

Oxidative pyrolysis

Small-UV

Large-UV

Well-V

Oxidative pyrolysis

Small-UV

Polystyrene

Large-UV

Well-V

Large-UV

Oxidative pyrolysis

Small-UV

Well-V

0.0

849

850

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

is >1, indicating the lethality of the diluted fire atmosphere over 30 min. PVC is one of the few materials to show a predicted combustion toxicity that is almost independent of the equivalence ratio, showing unusually high toxicity under well-ventilated conditions. As an irritant gas, HCl will have the greatest effect of any toxic species. For underventilated fires, HCN from PA 6.6 makes the most significant contribution to the toxicity, and a high dependency on fire conditions is observed. HCN generated during small and large underventilated flaming tests makes the most significant contribution to the toxicity. For PA 6.6 in underventilated conditions, and for PVC, HCN, and HCl make the greatest contribution to the final FED, showing the importance of two toxicants other than CO. PMMA, LDPE, and PS are hydrocarbons without any halogens or nitrogen, and this is reflected in FED values well below 1 for the well-ventilated fire scenario. The fire effluent from PS has a relatively low toxicity, and shows a characteristically low dependence on fire conditions. The toxic contribution of CO under fuel-rich conditions is remarkably similar to those generated under fuel lean conditions, presented in section 2.2.2.

2.2 Effect of fire retardants and nanofillers on fire toxicity 2.2.1 Effect of fire retardants and nanoclays on fire toxicity of polypropylene Samples of PP containing 5% PP grafted with maleic anhydride, to allow dispersion of nanoclay (NC), were compounded with 30% ammonium polyphosphate fire retardant and/or 5% Cloisite 20A NC. The toxicity of the PP materials is much lower than that for the PA samples in underventilated conditions (Figure 8). Again there is a progressive increase in the toxicity from small underventilated to large underventilated, which is shown consistently across the samples. Although the addition of either fire retardant (FR) or NC has no significant effect on the CO yield, it appears that the combination of both NC and FR results in higher CO yields for underventilated flaming [37,38].

2.2.2 Effect of fire retardants on ethylene-vinyl acetate copolymer Ethylene-vinyl acetate copolymer (EVA) is widely used with mineral fillers in the cable industry, as a material for avoiding the toxic, smoky, corrosive effluent of PVC-sheathed cables. EVA containing 27% vinyl acetate was compounded into low smoke and fume formulation. Fire retarded composites were formulated with 30% by weight of the EVA and 70% aluminum hydroxide (ATH) or 65% ATH and either 5% zinc hydroxystannate (ZS) (ZnSn(OH)6), 5% magnesium borate (MgB) (MgO (B2O3)2 H2O), or 5% zinc borate (ZB) (2ZnO.3B2O3.3.5H2O) [34]. In order to compare the effect of fillers on the combustion chemistry of composite EVA materials, the same mass of EVA was used for each experiment; thus, the filled composites containing 30% polymer and 70% fire retardant filler had a sample mass 3.33 times greater than those of pure EVA [35]. The contributions to FED, shown in Figure 9, indicate that except as a diluent filler, ATH and ATH with MgB or ZS have only a modest effect on fire toxicity,

2. Analysis of fire toxicity for different polymers and products

2.5

CO

Hypoxia

Organics

FED

2.0

1.5

1.0

0.5

PP

PP+FR

PP+NC

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

0.0

PP+FR+NC

FIGURE 8 Fractional effective dose (FED) for polypropylene (PP) with fire retardant and nanoclays (NCs) [37,38].

increasing the CO contribution during underventilated burning compared with EVA alone. This is believed [34] to result from catalytic oxidation of char (after-glow) by the freshly formed alumina surface. In contrast, ATH with ZB has a dramatic effect on reducing the carbon monoxide contribution to fire toxicity especially in underventilated conditions. It is thought that ZB forms a glass, which destroys the charoxidizing catalytic properties of the freshly formed alumina, leaving more of the carbonaceous residue in the condensed phase [35]. A separate study, using different EVA, with 68% ATH or magnesium hydroxide, or 63% when used with 5% NC (Figure 10) shows that ATH and magnesium hydroxide enhance the CO yield compared to the pure polymer, or polymer plus NC, which had little effect [37].

2.2.3 Effect of fire retardants on polyester materials Polybutylene terephthalate was fire retarded with 18% aluminum phosphinate Exolit OP1240 and/or 5% of Cloisite 30B or sepiolite NC. Figure 11 shows the consistently higher toxicity from burning PBT under different fire conditions and the additional increase in toxicity resulting from the use of fire retardant presumably through gas phase inhibition. It is interesting to see this effect disappear on incorporation of either sepiolite or cloisite. A slight increase in CO yield from well-ventilated to underventilated conditions for PBT and PBT with sepiolite, and PBT þ FR combinations is observed. As aluminum phosphinate is a gas phase flame retardant, the

851

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

1.0

CO

Hypoxia

FED

0.8

0.6

0.4

0.2

0.0 Well-V Under-V Well-V

EVA

UV

Well-V

EVA/ATH

UV

Well-V

EVA/ATH/MgB

UV

Well-V

EVA/ATH/ZB

UV

EVA/ATH/ ZS

FIGURE 9 Contribution of each component to toxicity at different ventilation conditions for (ethylenevinyl acetate copolymer) EVA fire retarded composites [2,35].

1.4

CO

Organic

Hypoxia

1.2 1.0

FED

0.8 0.6 0.4 0.2

EVA

EVA+ATH

Large-UV

Well-V

Small-UV

Large-UV

EVA+MgOH+NC

Oxidative pyrolysis

Well-V

Small-UV

Oxidative pyrolysis

Small-UV

EVA+NC

Large-UV

Well-V

Large-UV

EVA+MgOH

Oxidative pyrolysis

Well-V

Small-UV

Large-UV

Oxidative pyrolysis

Well-V

Small-UV

Large-UV

Oxidative pyrolysis

Well-V

Small-UV

0.0 Oxidative pyrolysis

852

EVA+ATH+NC

FIGURE 10 Contributions to (fractional effective dose) FED for ethylene-vinyl acetate copolymer (EVA) with fire retardant and nanoclay (NC) [37].

2. Analysis of fire toxicity for different polymers and products

CO

Hypoxia

Large-UV

1.2

Well-V

1.4 Organic

FED

1.0 0.8 0.6 0.4 0.2

PBT

PBT+FR

PBT+Sep

PBT+Clo

Large-UV

Well-V

Small-UV

Large-UV

PBT+FR+Clo

Oxidative pyrolysis

Well-V

Small-UV

Large-UV

Oxidative pyrolysis

Well-V

Small-UV

Oxidative pyrolysis

Small-UV

Large-UV

Oxidative pyrolysis

Well-V

Small-UV

Large-UV

Oxidative pyrolysis

Well-V

Small-UV

Oxidative pyrolysis

0.0

PBT+FR+Sep

FIGURE 11 Contributions to (fractional effective dose) FED for (persistent bioaccumulative and toxic) PBT with fire retardant and nanoclays (NCs) [2,37].

increase with FR would be expected, but this effect seems to diminish in the presence of the NCs. The NC alone has little effect on the CO yield [37]. In a separate study (Figure 12), fire retarded glass-reinforced polyester (GRP) roof light material, supplied as a translucent flat sheet formulated to achieve a “class 1” rating to BS476 Part 7 was investigated [39]. This material contained both chlorine and bromine. The major effect on the fire toxicity of the presence of HCl and HBr was to increase the yield of carbon monoxide increasing its contribution to the toxicity, especially in well-ventilated combustion. CO makes the most significant contribution to the toxicity under all conditions, contributing around 80% of the FED. The CO contribution to the FED is hardly changed in different flaming scenarios, resulting in FEDs with a remarkably low sensitivity to fire conditions [2].

2.2.4 Effect of fire retardants and NCs on PAs Samples of PA 6 also containing 18% aluminum phosphinate melamine polyphosphate mixture (Clariant’s OP 1311) and/or 5% Cloisite 30B NC shared modest reduction in fire toxicity when compared to the virgin polymer. For oxidative pyrolysis, the only significant contribution to the toxicity comes from NO2. The total value of the FED is higher for well-ventilated, small and large underventilated flaming. Similar to PA 6.6, the most toxicologically significant product from PA 6 is hydrogen cyanide (Figure 13). As the ventilation becomes limited, both the CO and HCN contributions increase while the oxygen depletion, by the nature of the experiment, decreases. In addition, a relationship between hypoxia and

853

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

3.0 CO

Hypoxia

HCl

HBr

Organic

FED

2.0

1.0

0.0

Oxidative pyrolysis

Well-V

Small UV

Large UV

FIGURE 12 Contributions to (fractional effective dose) FED for (glass-reinforced polyester) GRP [2].

10.0 CO

HCN

Hypoxia

NO2

Organics

8.0

FED

6.0

4.0

2.0

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

PA6+NC

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

PA6+FR

Large-UV

Small-UV

Well-V

Oxidative pyrolysis

PA6

Large-UV

Small-UV

Well-V

0.0

Oxidative pyrolysis

854

PA6+FR+NC

FIGURE 13 Contributions to fractional effective dose (FED) for polyamide (PA) 6 with fire retardant and nanoclay (NC) [2,37].

2. Analysis of fire toxicity for different polymers and products

unburned hydrocarbons has been observed, as the production of both species is favored by underventilated conditions. The increase of HCN concentration and organic species, and CO and oxygen depletion are independent of the temperature of the furnace and give higher yields with reduced ventilation [38,40]. For wellventilated burning, where the major toxicant is NO2, there are only minor differences between the FR and NC materials compared to that for the virgin polymer. There is a decrease in CO for all the materials in the most dangerous, underventilated conditions on incorporation of either NC or FR, and only a slight increase when both are present [2,37]. Samples of PA 6, formulated to industry standards to meet UL94 V-0 at 0.8 mm, containing 30% glass fiber (abbreviated to PA6 þ GF), for electrical applications, and fire retarded with either brominated PS and antimony trioxide (20%) (abbreviated to BrSb), or a blend of aluminum phosphinate and melamine polyphosphate (AlPiM), were burned under well-ventilated flaming conditions [41]. The relative contribution to the toxicity of individual components using FED, based on the Purser model is presented in Figure 14. Figure 14 shows estimates of toxicity expressed as FED (assuming a loading of 1 kg of material in 50 m3). They show a slight change in toxicity resulting from the presence of AlPiM (more HCN, but less NOx), but a dramatic increase in both CO and HCN yields. It is notable that the fire toxicity of the PA 6 þ GF þ BrSb materials is a factor of 10 larger than that of PA 6 þ GF and a factor of 5 larger than the

10.0 CO

HCN

HBr

NO2

0.56

0.47

0.54

0.45

Hypoxia

FED

8.0

6.0

4.0

2.0

0.0 0.54

650°C

825°C

PA 6 + GF

0.49

650°C

0.49

0.51

825°C

PA 6 + GF + AlPiM

0.60

0.98

650°C

0.67

1.00

825°C

PA 6 + GF + BrSb

FIGURE 14 Fractional effective dose (FED) comparison for two different flame retardant systems in PA6 containing 30% glass fiber [42].

855

856

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

PA6 þ GF þ AlPiM material at 650  C; at 825  C it is a factor of 30 larger than the PA 6 þ GF, and 17 larger than the AlPiM material. The BrSb materials all show high fire toxicity, even under the normally least toxic, well-ventilated fire conditions. The lower fire toxicity of the AlPiM materials at higher temperatures suggests a reduction in fire hazard. It appears that the AlPiM system “switches off” this flame quenching process as the fire becomes more severe, reducing the toxicity of the effluent.

2.3 Quantification of effluent toxicity during burning of fire retarded commercial products (precise formulation unknown) 2.3.1 Medium density fiber board Medium density fiber board (MDF) is a wood based composite, composed of around 50% cellulose, 25% hemicellulose, and 25% lignin. MDF is a generic name used to define sheets composed mainly of lignin-cellulosic fibers glued with a synthetic resin or some other suitable adhesive system, combined together by means of pressure and heat. Samples of commercial MDF without and with flame retardant (MDF-FR) were investigated for fire toxicity. MDF was a BS476 Part 7 “class 2” rated non-flame-retarded construction material. Fire-resistant MDF (MDF-FR) was an MDF construction material treated with a fire retardant, which contained nitrogen, bromine, and phosphorus. The product met BS476 Part 7 “class 1” rated MDF. Both materials had a nominal density of 740 kg m3. Although the relative yields are similar to LDPE and PS, the CO2 yield from MDF under well-ventilated combustion conditions is considerably lower, corresponding to the lower carbon and higher oxygen content. The combustion is also more efficient than for the CH polymers, perhaps due to the additional oxygen in the molecular structure. Another important feature of cellulosic materials is that they are char formers, so that, especially in underventilated conditions, a significant residue remains, most of which is carbon. The yield of organic carbon is less than for the CH polymers, but the conversion of fuel carbon to CO is greater. MDF also contains approximately 3% nitrogen presumably from the glue or fire retardant, so a small amount of HCN is produced, especially in underventilated conditions. Small quantities of hydrogen chloride and hydrogen bromide were generated from MDF combustion and MDF-FR. The toxic product yields obtained from the MDF and MDF-FR combustion under different fire conditions are presented in Figure 15. There are significant differences between MDF and MDF-FR. For both furnace temperatures for MDF, CO, and HCN yields increase with the increase of equivalence ratio, while NOx yields decrease. However, organics contribute more to the FED. A different pattern is observed in the case of MDF-FR, which could be explained by the presence of the brominated flame retardant, which increases the CO, HCN, and NOx yields. For MDF-FR, the most significant contributions to the toxicity are NOx and HCN. These differences are the greatest at the higher furnace temperature. However, they seem to be

2. Analysis of fire toxicity for different polymers and products

4.0 CO

HCN

NOx

HBr/HCl

Hypoxia

Organics

FED

3.0

2.0

1.0

MDF

Large-UV

Small-UV

Well-V

Large-UV

Small-UV

Well-V

0.0

MDF-FR

FIGURE 15 Contributions to fractional effective dose (FED) comparison for medium density fiber (MDF) and medium density fiber-flame retardant (MDF-FR).

independent of the temperature and ventilation conditions. Overall, the predicted toxicity is higher for MDF-FR than for MDF.

2.3.2 Insulation products Modern, lightweight building materials are cheaper to produce, transport, and erect, and offer improved thermal insulation, allowing more efficient temperature control. In comparison to traditional materials, many insulation materials present a greater fire hazard, being less effective fire barriers, more combustible and having higher fire toxicity. Most common insulation materials comprise gases trapped in a matrix to inhibit convective heat transfer. The fire toxicity of six insulation materials: glass wool (GW), stone wool (SW), expanded PS foam (EPS), phenolic foam (PhF), PUR, and polyisocyanurate foam (PIR) were investigated under a range of fire conditions, oxidative pyrolysis, and well-ventilated flaming to underventilated flaming [42]. Both GW and SW are classified as noncombustible or limited combustibility depending on the binder content. While both lose small (w5%) quantities of pyrolyzable binders, most of the mass will not burn, and there is insufficient fuel for a flame to propagate through the bulk of the material, so their contribution to the fuel load is negligible. The foamed materials are organic polymer based, and depending on the material and fire conditions, a significant part of their mass is lost as fuel, and may contribute to the overall size of the fire. The thermal insulation materials were selected to have comparable densities [43].

857

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

For the two fibrous materials in the flaming condition, SW and GW, ignition was not observed even above 800  C (nf). For the four foams, PUR, PIR, PhF, and EPS, ignition and steady flaming was achieved for the two flaming fire conditions and for intermediate ventilation conditions [42]. Figure 16 shows the increase in the carbon monoxide yield as the ventilation changes from well-ventilated to underventilated. In comparison to polymers without flame retardants, the CO yields for PIR and PUR in well-ventilated conditions are high. This suggests the presence of gas phase free radical quenchers, such as halogens or volatile phosphorus compounds, preventing the conversion of CO to CO2 by reducing the availability of the OH$ radical [36]. For the two nitrogen-containing polymers, PUR and PIR, the HCN yield is significant, both in well-ventilated flaming, and for underventilated flaming, and those materials have much greater fire toxicity than EPS or PhF. The GW and SW products show very low fire toxicity. The variation of NO2 with fire conditions for the three materials shows less consistent trends although the PhF seems to show a progressive increase as the fire condition becomes underventilated. For PIR and PUR, there is a slight decreasing trend with underventilation, which corresponds to the increased yields of HCN, and reduced availability of oxygen. HCl and HBr were close to limits of detection.

2.3.2.1 Isocyanates from insulation materials Although not specifically included in the normal lists of fire effluents for quantification of fire toxicity, it is been suggested that isocyanates (molecules with functional group eNCO, used in PUs and some binders) may pose a hitherto unquantified

4.0

3.0

FED

CO

HCN

Hypoxia

HCl

NO2

HBr

2.0

1.0

PIR

PUR

PHF

EPS

SW

T = 825C nf

Oxidative pyrolysis

T = 825C nf

Oxidative pyrolysis

Large UV

Well-V

Small UV

Oxidative pyrolysis

Large UV

Well-V

Small UV

Oxidative pyrolysis

Large UV

Well-V

Small UV

Oxidative pyrolysis

Large UV

Well-V

Small UV

0.0 Oxidative pyrolysis

858

GW

FIGURE 16 Fractional effective dose (FED) for insulation materials (for oxidative pyrolysis and flaming conditions, except stone wool (SW) and glass wool (GW) nonflaming (nf)) [42].

2. Analysis of fire toxicity for different polymers and products

hazard in fire effluents. In ISO TC92 SC3, the main fire toxicants identified do not include isocyanates, although there is some evidence that they are present in fire effluents in toxicologicially significant concentrations. A cone calorimeter study [44] included five insulation materials, GW, SW, EPS foam, PUR foam, and PIR foam, as part of a larger project to investigate the presence of isocyanates in fire effluents. Each sample was exposed to an intermediate heat flux of 35 kW m2. It was found that isocyanate production was favored in the early well-ventilated stages of flaming, while hydrogen cyanide was favored in the more toxicologically significant underventilated stages [42]. For noncombustible samples, GW and SW, the materials were subject to a full 15-min pyrolysis, during which isocyanate collection continued. PS foam was only subjected to 10-min pyrolysis during which it did not ignite, while the PUR and PIR foams burnt for 5.5 and 9 min, respectively [45]. The isocyanates were only collected during these times. Thus, the sample collection time was greatest for the least flammable samples. From the reported data, isocyanate yields have been calculated as shown in Table 3. In the original paper, the mass loss data for PUR or PIR is not givendfor the calculation, data from experiments reported elsewhere [42] have been used. Despite the 37.5% mass loss from PS, no data are presented on the composition of the volatile products, and no explanation is provided, but it may be that a large amount of soot blocked the sampling lines. The yields are calculated on a mass charge basis (favored by engineers as it indicates the total amount of product that may be formed per unit mass present in a building) and as originally reported, on a mass loss basis (materials such as GW which were found to be 88.5% nonvolatile (glass fiber) and 11.5% organic binder), which is the yield from the organic binder alone, as though there was no glass fiber present [42]. The use of the mass loss yield results in a misleading impression of large isocyanate yields from the inorganic fiber insulation materials. In effect, it compares the yield from 1 kg of PUR foam with that from 9 kg of GW insulation, despite the fact

Table 3 Isocyanate Yield and Calculation Data [42]

GW SW PS PUR PIR

Concentration mg mL3

Mass of Sample/g

Mass Loss/g

Mass of Isocyanate mgL1

8100 990 0 4500 3350

20 14 8 17 14

2.3 0.5 3.0 13.0 11.5

175 21 0 36 43

Mass Charge Yield mg gL1

Mass Loss Yield mg gL1

8.7 1.5 0 2.1 3.1

82.1 68.9 0 2.7 7

GW, glass wool; PS, polystyrene; PUR, polyurethane foam; PIR, polyisocyanurate foam; SW, stone wool.

859

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

that the latter is much less flammable, and would therefore only be volatilized in the presence of a much stronger external heat source [42].

2.3.3 PVC carpet Another commercial material that has been investigated is PVC carpet, commonly found in domestic buildings [28,46,47]. The PVC carpet had a thickness of 2.0 mm, and contained almost as much plasticizer and mineral fillers as PVC polymer. The contribution of carbon monoxide and hydrogen chloride to FED is shown in Figure 17. For all fire conditions, HCl makes the greatest contribution to the toxicity. In common with the other halogenated materials investigated, increased yields for CO are observed. The HCl yields found from well-ventilated and underventilated combustion were close to the theoretical maximum. This supports an earlier study where the HCl yields are independent of fire stages and temperature effects [17,36].

2.3.3.1 Particulates, PAHs, PCDD/Fs, isocyanates from PVC carpet The PVC carpet was subjected to an in-depth study to look for other noxious substances in addition to the eight standard ISO fire gases. It was found that in wellventilated and underventilated tests on PVC carpet both large and small particles are generated (of the order of 1e4 mm), whereas in oxidative pyrolysis, only a narrow distribution of large particle sizes was found. Compared to well-ventilated tests, greater quantities of smaller, more dangerous particle sizes were found for the underventilated fire scenario. As small particles have a higher proportion of surface area

4.0 CO

HCl

Hypoxia

3.0 FED

860

2.0

1.0

0.0

Oxidative pyrolysis

Well-ventilated

Under-ventilated

FIGURE 17 Contributions to Fractional effective dose (FED) for polyvinyl chloride (PVC) carpet [47].

2. Analysis of fire toxicity for different polymers and products

400 Oxidative pyrolysis 300

Well-ventilated Under-ventilated

200 100

OCDF

1,2,3,4,7,8,9-HpCDF

1,2,3,4,6,7,8-HpCDF

2,3,4,6,7,8-HxCDF

1,2,3,6,7,8-HxCDF

1,2,3,7,8,9-HxCDF

2,3,4,7,8-PCDF

1,2,3,4,7,8-HxCDF

1,2,3,7,8-PCDF

2,3,7,8-TCDF

OCDD

1,2,3,4,6,7,8-HpCDD

1,2,3,6,7,8-HxCDD

1,2,3,7,8,9-HxCDD

1,2,3,4,7,8-HxCDD

2,3,7,8-TCDD

0

1,2,3,7,8-PCDD

PCDD/F emission, pg g-1 residue

per mass than do larger particles, it is possible that more toxic substances may be adsorbed onto smaller particles [28,46,47]. Although some chlorine will always appear as HCl, other chlorine containing gas or vapor species will be produced while, in some formulations, some chlorine may remain in the residue. A number of chlorine-containing species have been identified from large-scale fires burning a high proportion of PVC, including monochlorobenzene and dichlorobenzene and other chloroaromatic and chloroaliphatic hydrocarbons. Evidence exists to show that, depending on the fire situation, as much as 20% of the chlorine may exist in an organic form [48]. It was also found that the amount of chlorine from HCl could be very high in soot, even though the major part of the HCl produced was shown to be present in the gas phase [46,47]. In this case, for the PVC carpet, it is reported that the chlorine content adsorbed on the soot was in the range between 0.5 and 2.5 wt%, while in nonflaming fire conditions around approximately 18e20 wt% of chlorine is present in residue, compared to 10 wt% for flaming conditions which favor polychlorodibenzo furan and dioxin formation. These highly toxic materials are usually qualified by their toxic equivalence to 2,3,7,8-tetrachlorodibenzodioxin as toxic equivalency (TEQ). Total amounts of PCDD/F in the residue from PVC carpet shown in Figure 18 varies for different fire scenarios between 11 and 394 pg International toxic equivalents per gram (ITEQ/g) of burned material, corresponding to 41e1170 pg ITEQ/g residue [28]. From the studies, it is interesting to note that both well-ventilated and underventilated combustion of the PVC carpet produced the complete range of PAHs. In the

FIGURE 18 Concentration of polychloro- and polybromo-dibenzo-dioxins and furans (PCDD/F) in polyvinyl chloride (PVC) carpet residues under different fire conditions.

861

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

pyrolysis tests, however, only volatile and semivolatile PAHs were produced. For both well-ventilated and underventilated conditions, the toxicity weighted yields for the associated particle PAHs increased and generally dominated over the volatile species. The toxicity weighted yields for the underventilated tests were orders of magnitude higher than the yields from the well-ventilated tests, whereas for oxidative pyrolysis conditions, it is observed that the volatile part dominated the toxicity [46,47].

2.4 Cables and cable materials Ten commercial cables (Figure 19) were investigated for fire toxicity using the conditions specified in the precursor International Electrotechnical Commission (IEC) standard, IEC 60695-7-50 [49], which uses the steady-state tube furnace, but a simplistic methodology for ensuring the setting up of the tube furnace ventilation condition to obtain a particular fire stage. For each cable, the FED is shown for each of three fire stages. In most cases (8/10), oxidative pyrolysis is the least toxic, presumably due to the low mass loss of the polyolefin polymers at 350  C [50]. However, this temperature is high enough to release HCl from PVC in two cases. In general, the FED, based on the limited range of gases analyzed, increases from oxidative pyrolysis, to well-ventilated, to developed flaming. In common with most materials, the most hazardous fire condition is confirmed to be the underventilated fully developed fire [2].

2.0

CO

HCl

Hypoxia

1.5

FED

1.0

0.5

2 Halogen free data cable

3

2

3

Halogen free power cables

FIGURE 19 Contribution of individual gases to toxicity from burning cables.

Well-V

Large-UV

Oxidative pyrolysis

Well-V

Large-UV

Oxidative pyrolysis

Well-V 1

Large-UV

Oxidative pyrolysis

Well-V

Large-UV

Oxidative pyrolysis

Well-V

Large-UV

Oxidative pyrolysis

Well-V 1

PVC data cables

Large-UV

Oxidative pyrolysis

Well-V

Large-UV

Oxidative pyrolysis

Well-V

Large-UV

Oxidative pyrolysis

Well-V

Large-UV

Oxidative pyrolysis

Well-V 1

Large-UV

0.0 Oxidative pyrolysis

862

2

PVC power cables

PVC single conductor cable

2. Analysis of fire toxicity for different polymers and products

The two major toxicants are seen to be CO and HCl. However, it is notable that the CO yield the PVC fires is greater than that in halogen-free fires, and it is also noted that the CO yield in PVC fires increases with underventilation, whereas for pure PVC, the yield is almost independent of ventilation [2].

3. CONCLUSIONS Fire gases contain a mixture of fully oxidized products, such as CO2, partially oxidized products, such as CO and aldehydes, fuel, or fuel degradation products, including aliphatic or aromatic hydrocarbons, and other stable gas molecules, such as nitrogen and hydrogen halides. Most fire deaths arise from the inhalation of toxic gases, though the incapacitating effects of smoke and irritant gases also play an important role. Fire growth has been classified into a number of stages, from oxidative pyrolysis and early wellventilated flaming, through to fully developed underventilated flaming. For many materials (such as CH- and CHO-containing polymers), the yield of the main toxicant, CO, has been shown to depend only on fire conditions, not on the nature of the polymer. It is therefore essential to the assessment of fire gas toxicity that these different fire stages can be adequately replicated. The steady-state tube furnace (ISO 19700) provides an excellent method for exploring the relationship between combustion conditions and product yields. The yields of most of these species will depend on the material, the decomposition conditions (nonflaming or flaming), and for flaming, the ventilation conditions. For nonflaming decomposition conditions, some examples are presented of the effects of temperature and air flow on product yields, while flaming combustion conditions are expressed in terms of the fuel/air equivalence ratio, f. For flaming conditions, examples are presented for five different natural and synthetic polymer classes demonstrating the relationship between the equivalence ratio and the product yields. The main carbon-containing products CO2, CO, soot particulates (as total organic carbon in the gas phase and is expressed as the yield of CH2) are reported. In addition, the yields of nitrogen-containing products (HCN, NO, and NO2), and the HCl yield from a material with a high chlorine content (PVC), are presented. Yields for all gas phase products and residues are expressed as grams per gram (i.e. grams of product per gram of material on a mass charge basis). Equivalence ratios are varied from 0.5 to 2.5 in separate experiments representing wellventilated to underventilated flaming. For most natural and synthetic polymers, the yields of toxic products such as carbon monoxide are very low under well-ventilated combustion conditions, but increase steeply in underventilated combustion conditions as f increases above 1. This results in a sigmoid relationship between equivalence ratio and yield, so that the yields of products of incomplete combustion can increase by factors of approximately 50 between well-ventilated flaming conditions (f w0.7), and fuel-rich conditions (f w1.5). The increase in products such as CO, HCN, organic irritants,

863

864

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

and smoke particulates with 4 is matched by a corresponding decrease in the products of more complete combustiondprincipally CO2, H2O, N2, and NOx. Superimposed on this basic pattern of decomposition and yields are lesser effects of decomposition temperature and oxygen concentration and specific effects related to the polymer composition. One important composition-related effect is residue formation in some natural and synthetic polymers, and in products containing fire retardants acting in the solid phase (such as phosphates), which leads to reduced yields of airborne products and particular trapping of fuel carbon. Other important yield modifiers are halogenated fire retardants, which act in the gas phase, resulting in increased yields of products of incomplete combustion particularly carbon monoxide and hydrogen cyanide throughout the range of equivalence ratios. Assessment of fire gas toxicity is an essential component of fire hazard analysis. The toxicities of the effluents, showing the contribution of individual toxic components, are compared using the FED model. The predicted toxicities show variations of up to two orders of magnitude with change in fire scenario. They also show changes of at least one order of magnitude for different materials in the same fire scenario. Finally, they show that in many cases CO, which is often assumed to be the most, or even the only toxicologically significant fire gas, is of less importance than HCl or HCN when chlorine or nitrogen is present in the compound. For polyisocyanurate and PUR, there is a significant contribution from hydrogen cyanide resulting in a doubling of the overall toxicity, as the fire condition changes from well-ventilated to underventilated. The presence of fire retardants and the incorporation of NC reduce the flammability, but for many fire retardants, there is no significant adverse effect of these additives on the toxicity of the material studies under most fire conditions. However, higher yields of CO under well-ventilated fire scenarios are often observed. The insulation materials showed an order of increasing fire toxicity, from SW (least toxic), GW, PS, phenolic, PU to polyisocyanurate foam (most toxic). The current work shows lower carbon monoxide yields for all materials under well-ventilated conditions, compared to underventilated conditions, although the presence of halogens (presumably present as flame retardants) increases the CO and HCN yields in well-ventilated conditions.

References [1] Fire Statistics United Kingdom 2008. Communities and local government; 2012. [2] Stec AA, Hull TR, editors. Fire toxicity. Cambridge: Woodhead Publishing; 2010. [3] ISO/TR 16738. Fire-safety engineeringdtechnical information on methods for evaluating behaviour and movement of people; 2009. [4] Purser DA. Assessment of hazards to occupants from smoke, toxic gases and heat. In: DiNenno PJ, editor. Chapter 2e6, SFPE handbook of fire protection engineering. 4th ed. Quincy (MA, USA): National Fire Protection Association; 2008. pp. 2-96e2-193.

References

[5] Stec AA, Hull TR, Lebek K, Purser JA, Purser DA. The effect of temperature and ventilation condition on the toxic product yields from burning polymers. Fire Mater 2008;32: 49e60. [6] ISO TS 19706. Guidelines for assessing the fire threat to people; 2011. [7] Purser DA. Toxic product yields and hazard assessment for fully enclosed design fires. Polym Int 2000;49:1232e55. [8] Pitts WM. The global equivalence ratio concept and the formation mechanisms of carbon monoxide in enclosure fires. Prog Energy Combust Sci 1995;21:197e237. [9] Hull TR, Paul KT. Bench-scale assessment of combustion toxicityda critical analysis of current protocols. Fire Saf J 2007;42:340e65. [10] Hull TR, Stec AA, Lebek K, Price D. Factors affecting the combustion toxicity of polymeric materials. Polym Degrad Stab 2007;92:2239e46. [11] Stec AA, Hull TR. Fire Toxicity Assessment: Comparison of Asphyxiant Yields from Laboratory and Large Scale Flaming Fires Fire Safety Science, (in press 2014). [12] ISO 9705. Fire testsdfull scale room test for surface products; 1993. [13] Blomqvist P, Lonnermark A. Characterization of the combustion products in largescale fire tests: comparison of three experimental configurations. Fire Mater 2001; 25:71e81. [14] Stec AA, Hull TR, Purser JA, Purser DA. Comparison of toxic product yields from bench-scale to ISO room. Fire Saf J 2009;44:62e70. [15] ISO TS 19700. Controlled equivalence ratio method for the determination of hazardous components of fire effluents; 2007. [16] BS 7990. Tube furnace method for the determination of toxic product yields in fire effluents; 2003. [17] Stec AA, Hull TR, Purser JA, Blomqvist P, Lebek K. A comparison of toxic product yields obtained from five laboratories using the steady state tube furnace (ISO TS 19700). Fire Saf Sci 2008:653e64. [18] Purser DA. The evolution of toxic effluents in fires and the assessment of toxic hazard. Toxicol Lett 1992;64-65:247e55. [19] Stec AA, Hull TR, Lebek K. Characterisation of the steady state tube furnace (ISO TS 19700) for fire toxicity assessment. Polym Degrad Stab 2008;93:2058e65. [20] Blomqvist P, Hertzberg T, Tuovinen H, Arrhenius K, Rosell L. Detailed determination of smoke gas contents using a small-scale controlled equivalence ratio tube furnace method. Fire Mater 2007;31:495e521. [21] Blomqvist P, Hertzberg T, Tuovinen H. A small-scale controlled equivalence ratio tube furnace methoddexperience of the method and the link to large scale fires. In: Proceedings of the 10th international fire science and engineering conference (Interflam), Edinburgh; 2007. [22] Hull TR, Lebek K, Stec AA, Paul KT, Price D, Schartel B. In: Bench-scale assessment of fire toxicity, advances in the flame retardancy of polymeric materials: current perspectives presented at FRPM’05. Norderstedt: Herstellung und Verlag; 2007. pp. 235e48. [23] Hull TR, Lebek K, Pezzani M, Messa S. Comparison of toxic product yields of burning cables in bench and large-scale experiments. Fire Saf J 2008;43:140e50. [24] ISO 13344. Estimation of the lethal toxic potency of fire effluents; 2004. [25] ISO 13571. Life threat from fires e guidance on the estimation of time available for escape using fire data; 2012.

865

866

CHAPTER 24 Influence of Fire Retardants and Nanofillers on Fire Toxicity

[26] Rhodes J, Smith C, Stec AA. Characterisation of soot particulates from fire retarded and nanocomposite materials, and their toxicological impact. Polym Degrad Stab 2011;96: 277e84. [27] WHO, World Health Organisation. Protection of the human environment, outdoor air pollution. Geneva; 2004. [28] Stec AA, Readman J, Blomqvist P, Gylestam D, Karlsson D, Wojtalewicz D, et al. Analysis of toxic effluents released from PVC carpet under different fire conditions. Chemosphere 2013;90:65e71. [29] Huang H, Buekens A. On the mechanisms of dioxin formation in combustion processes. Chemosphere 1995;31:4099e117. [30] Blomqvist P, Rosell L, Simonson M. Emissions from fires part II: simulated room fires. Fire Technol 2004;40:59e73. [31] ISO 26367-1. Guidelines for assessing the adverse environmental impact of fire effluentsdpart I: General, 2011; 2011. [32] Stec AA, Hull TR, Yields. Effects from common materials under different fire conditions. In: Bradley D, editor. Proceedings of the 6th international seminar on fire and explosion hazards, Leeds, UK; 2010. pp. 782e91. [33] Purser DA. Influence of fire retardants on toxic and environmental hazards from fires [Chapter 24]. In: Hull TR, Kandola BK, editors. Fire retardancy of polymersdnew strategies and mechanisms. RSC Publishing; 2009. pp. 381e404. [34] Hull TR, Quinn RE, Areri IG, Purser DA. Combustion toxicity of fire retarded EVA. Polym Degrad Stab 2002;77:235e42. [35] Hull TR, Carman JM, Purser DA. Prediction of CO evolution from small-scale polymer fires. Polym Int 2000;49:1259e65. [36] Hull TR, Stec AA, Paul KT. Hydrogen chloride in fires. Fire Saf Sci 2008:665e76. [37] Stec AA, Hull TR, Torero JL, Carvel R, Rein G, Bourbigot S, et al. Effects of fire retardants and nanofillers on the fire toxicity, fire and polymers Vdmaterials and concepts for fire retardancy, Ch 21. In: ACS symposium series 1013. Oxford University Press; 2009. pp. 342e66. [38] Stec AA, Hull TR. Assessment of fire toxicity from polymer nanocomposites, fire retardancy of polymers: new strategies and mechanisms [Chapter 25]. In: Hull TR, Kandola BK, editors. Cambridge (UK): Royal Society of Chemistry; 2009. pp. 405e18. [39] BS 476-7: Fire tests on building materials and structures - Method of test to determine the classification of the surface spread of flame of products”; 1997. [40] Stec AA, Hull TR. Yields of toxic combustion products from fire retarded nanocomposite polymers. In: Drysdale D, Bradley D, Molkov V, Carvel R, editors. Proceedings of the 5th international seminar on fire and explosion hazards, Edinburgh, UK, April 2007. University of Edinburgh; 2008. Prize winning papers, Section 2, pp. 125e35. [41] Molyneux S, Stec AA, Hull TR. The effect of gas phase flame retardants on fire effluent toxicity. Polymer Degradation and Stability (in Press) 2013. http://dx.doi.org/10.1016/j. polymdegradstab.2013.09.013. [42] Stec AA, Hull TR. Assessment of the fire toxicity of building insulation materials. Energy Build 2011;43:498e506. [43] Papadopoulos AM. State of the art in thermal insulation materials and aims for future developments. Energy Build 2005;37:77e86. [44] Blomqvist P, Hertzberg T, Dalene M, Skarping G. Isocyanates, aminoisocyanates and amines from firesda screening of common materials found in buildings. Fire Mater 2003;27:275e94.

References

[45] Hertzberg T, Blomqvist P, Dalene M, Skarping G. Particles and isocyanates from fires, SP report 2003:05. Boras: SP Swedish National Testing and Research Institute; 2003. [46] Blomqvist P, Simonson McNamee M, Stec A, Gylenstam D, Karlsson D. Characterisation of fire generated particles. Sp Rep 2010;01:2010. [47] Blomqvist P, Simonson McNamee M, Stec AA, Gylestam D, Karlsson D. Detailed study of distribution patterns of polycyclic aromatic hydrocarbons and isocyanates under different fire conditions. Fire and Materials 2014;38:125e44. [48] Lebek K, Hull TR, Price D. Products of rigid PVC burning under various fire conditions, fire and polymers: materials and concepts for hazard prevention. ACS Symposium Series No.922. Oxford University Press; 2005. pp. 334e47. [49] IEC TS 60695-7-50 Fire hazard testing Part 7-50: toxicity of fire effluent estimation of toxic potency apparatus and test method. [50] Hull TR, Lebek K, Robinson JE. Acidity, toxicity and European cable regulation, Transactions of the international wire and cable symposium (Trans IWCS); 2006.

867