Samples of materials normally utilised as part of under-roof insulation systems were delivered for the assessment of the toxicity of their combustion gases. The purpose of the evaluation was to compare the toxicity indices of the XPS products.

    One gram of each product was burned in a chamber with a volume of 1m³. The concentrations of certain specified gases were determined by means of colorimetric (Dräger) tubes. These concentrations were then used to calculate the quantities of gases given off by burning 100g of material in a cubic metre of air.

    The toxicity index is calculated from the summation of the ratios of these concentrations to the concentrations causing fatality to man after a 30-minute exposure time.

    With sample B, some traces of halogen-containing gases were found by setting up the ion displacement reaction between silver nitrate and chloride or bromide ions to form insoluble silver chloride or silver bromide.

Or

AgNO₃ + 2Br- ® AgBr₂ (s) + NO₃

    These traces indicated the presence of halogen containing combustion products but did not quantify the concentrations. Compared to the other combustion products the concentration of halogen containing compounds was perceived to be of less significance.

    The fire retardant additives play a lesser role in the toxicity of the materials but this was difficult to quantify with this type of test. The halogen-based fire retardant that are normally used in polymeric materials, after thermal breakdown, tend to remain in a form too complex for this rather simple test protocol to identify. Therefore, although these combustion products would have a negative effect on the toxicity of gases emitted during combustion, we feel that for the purposes of comparison that this fact is not significant.
  
    Having evaluated and compared the toxicity indices of the three materials we would like to emphasize that toxicity is only one of the aspects that should be considered when the fire safety of a particular roof insulation system in under scrutiny. In essence a material must burn of decomposed thermally for any toxic combustion products to be liberated. It is our view that an insulation system which does not limit flame spread would in fact pose much more of a toxicity problem that a system that does not support significant flame spread, even if the toxicity of the gases produced on a mass to mass comparison is perceived to be low. The toxicity indices as determined, should therefore, be used in conjunction with the results of any large-scale flame spread evaluations in order to have more insight into the total fire risk presented by a particular under-roof insulation system.