Insulation Second Issue January 2018 Building Fabric vs Building Contents Smoke Toxicity from Room Fires A WHITE PAPER Low Energy Low Carbon Buildings
Executive Summary This paper presents and evaluates two fire tests conducted on a furnished domestic room. The tests only differed in the type of insulation material used in the room walls. One was insulated with PIR insulation and one was insulated with rock mineral fibre insulation. The tests were conducted in order to assess the relative contributions of the insulation materials and the room contents to the spread of fire and toxic gas emissions. The test results with regard to heat release rate, smoke and toxic gas emissions show that the two insulation materials behave similarly during the fire, and the main threat for occupants comes from the room s combustible contents. The main release of toxic gas emissions came from the burning of, and consequent smoke emission from, the room s combustible contents. A sharp peak in the concentration of the toxic gases is reached when flash over occurs. At this stage, the concentrations of carbon monoxide and hydrogen cyanide briefly but clearly exceed the 30 minute LC50 lethal concentration. The contribution of the insulation materials is negligible since the plasterboard acts as a thermal barrier. This is expected for a typical room fire and demonstrates the need to look at the fire and smoke performance of whole build ups instead of individual construction products. Towards the end of the tests, lower level toxic gas emissions were observed from the involvement of the insulation products in the fire. This was far after substantial FED and FEC values had been reached due to the burning of the contents of the room. The results demonstrate that efforts to reduce deaths from fires in dwellings should not focus solely on the performance of combustible construction products in general, and combustible insulation in particular. Efforts to improve the fire safety of dwelling contents and furnishings and the use of effective fire warning systems may have a much greater impact on fire safety. This occurred significantly before the building envelope and insulation materials became involved in the fire. 2
Introduction It is acknowledged that smoke inhalation causes or at least contributes to the majority of fatalities in fires in domestic properties. It is known that modern domestic room fires can become fully developed, and flashover can occur after only a few minutes. If an occupant has not escaped by this time, then they would be unlikely to survive. Occupants in other rooms may also be exposed to related smoke while in situ or while evacuating a building, although without themselves suffering the effects of the heat from flashover (because they are separated from the original fire). Building fires, and the evaluation of the toxic hazards from them, are very complex. The risk of building occupants being affected by toxic fumes depends on several factors such as the availability of combustible material, stage of the fire, and conditions of the combustion process, etc. This overall complexity makes it difficult to evaluate the smoke toxicity of a single product, in particular a construction product. In addition, the generation of toxic combustion products is not simply a material property: the smoke produced by burning construction products, and the resulting hazard, are strongly dependent on the way a product is integrated into the building, the fire scenario (e.g. room size, temperature, ventilation), and exposure time. This has led international standardisation committees addressing fire safety of buildings to conclude that, so far, there is no suitable test method for assessing the contribution of individual construction products to the emission of toxic gases linked to real fires. Despite this conclusion by experts in the field, there remains a significant lobby, funded in large part by the mineral fibre insulation industry, that seeks to associate combustible construction products in general, and combustible insulation materials in particular, with public safety as regards fire and smoke in buildings. Because of this lobby, the question has been raised in recent European discussions about what the role of construction products is in the overall generation of toxic smoke. In order to shed light on this issue, PU Europe commissioned Warrington Fire Ghent to conduct comparative tests, under the same fire conditions, of two identically furnished domestic rooms, one insulated with PIR insulation and one insulated with rock mineral fibre insulation. The objective of this study was to assess the contribution of the building fabric versus that of the building contents to heat release rate, smoke production and toxicity of fire effluents, and to determine whether there was any significant difference between the contributions of the two differing insulation materials. 3
Testing The inner dimensions of the test room without insulation were in accordance with the dimensions given in ISO 9705 (3.6 m x 2.4 m x 2.4 m). The floor, ceiling and walls were made of cellular concrete. The walls, apart from the wall containing the door, were insulated. The only difference between the two tests was that in one test the insulation was PIR and in the other it was rock mineral fibre. Details for the insulation products are given in Table 1. In order to achieve a fair comparison, the insulation thicknesses differed (80 mm vs. 140 mm) so that the wall U values in both tests were identical. In order to achieve the same inner volume of the test room in the test with PIR insulation, before mounting the insulation system an additional 50 mm cellular concrete layer was constructed inside ISO 9705 dimensioned room. Both rooms were identically furnished as detailed in Appendix A. The chosen fire scenario for this comparative testing simulates a waste bin fire (propane burner, 30 kw as defined in ISO 9705) that ignites the curtain and spreads to the armchair (first 5 minutes of the test). The burner was therefore placed in a corner and the curtain was installed just above the burner. After 5 minutes the burner was turned off and the further development of the fire was observed and analysed. The time of ignition of the armchair was chosen to be the starting point of the analysis, to minimise variations in fire spread in the early stages of the fire (burning curtains). In both tests, the insulation products were placed between 50 mm wide wooden battens at 570 mm centres, which were mechanically fixed to the concrete wall. The insulation layer was lined with 12.5 mm plasterboard. The plasterboard lining was applied with no horizontal joints, and with vertical joints located on top of the wooden battens. A power socket was placed near to the main fire load, to create a realistic weak spot in the plasterboard lining. Test 1 Test 2 Insulation product Rock mineral fibre without facing PIR insulation boards with composite foil facing on both sides Reaction to fire classification according to EN 13501-1 for the product as placed on A1 E the market Thermal conductivity (W/m K) 0.035 0.022 Thickness (mm) 140 80 Table 1: Insulation products used in Test 1 and Test 2 4
Results Heat Release Rate Figure 1 shows a graph of heat release rate. For details on the method of measurement see Appendix B. The point in time when the chair ignited defines the start of the test. Flashover occurred 7:24 mins. (444 secs.) after ignition of the chair in the rock mineral fibre test, and 6:55 mins. (415 secs.) in the PIR test. Flashover was caused by burning of the contents of the room only. In both tests, a second peak of the heat release rate was observed after 13 15 mins. (780 900 secs.). Subsequently, the fire decayed. First cracks appeared in the gypsum plasterboard in both tests after about 20 mins. (1,200 secs.). During the decay phase there was a slightly less steep decay in the PIR test, however, the HRR curves of both tests were below 50 kw after approximately 50 mins. Figure 1: Heat release rate versus smoke transmission 5
Results Concentration of Gaseous Effluents The following effluent gases were analysed (for further details see Appendix B): l carbon monoxide (CO); l carbon dioxide (CO 2 ); l hydrogen cyanide (HCN); l formaldehyde (CHOH); l acrolein (C 3 H 4 O); l sulfur dioxide (SO 2 ); l hydrogen chloride (HCl); and l NOx (measured by the sum of N 2 O, NO and NO 2 ). Gas concentrations correlated well with heat release rate. The example curves in Figure 2 show that for CO no significant difference can be seen. For HCN only in a very late phase of the test (more than 20 mins. after the start of the test) there is a slight increase in concentration in the PIR test compared with the rock mineral fibre test. This phenomenon can be attributed to the PIR insulation being partly exposed to radiative heat and direct flame impingement from the room s interiors by this time. The formaldehyde and acrolein curves indicate that in the same late phase of the tests the insulation was also contributing. However, in these cases there is a slight increase in concentration in the rock mineral fibre test compared with the PIR test. These late stage increases show that the contribution of the insulation products occurred very late and after much higher concentrations were reached in both tests due to the burning room contents. Concentration of carbon monoxide (CO). The currently accepted 30 minute LC 50 concentration for CO is 5700 ppm. Concentrations briefly but clearly exceed the 30 minute LC 50 lethal concentration. Concentration of formaldehyde (CHOH). The currently accepted 30 minute LC 50 concentration for CHOH is 750 ppm. Concentrations lie well below the 30 minute LC 50 lethal concentration. Concentration of hydrogen cyanide (HCN). The currently accepted 30 minute LC 50 concentration for HCN is 165 ppm. Concentrations briefly but clearly exceed the 30 minute LC 50 lethal concentration. Concentration of acrolein (C 3 H 4 O). The currently accepted 30 minute LC 50 concentration for C 3 H 4 O is 150 ppm. Concentrations lie well below the 30 minute LC 50 lethal concentration. Figure 2: Comparison of gas concentrations Note: The 30 min LC 50 concentration is the concentration of combustion products causing the death of 50 percent of animals when exposed 30 minutes to the given concentration. 6
Toxicity In order to assess the relative toxicity of the gaseous effluents from both tests, the Fractional Effective Dose (FED) and Fractional Effective Concentration (FEC) were computed by Exova Warringtonfire (UK), according to ISO 13571: 2012. FED is the dose received at time t divided by effective dose to cause incapacitation or death, where dose = concentration x time. FEC is the ratio of the concentration of an irritant at a point in time, to the concentration expected to cause incapacitation or death. Note: Toxicity is only to a certain degree a material property. It is strongly influenced by the environment, availability of oxygen, thermal attack, air flow and surfaces available for combustion. The chemistry of the combustion of a given material can therefore proceed along various routes and produce species in very different quantities dependant on conditions to which it is subjected. Such changes would impact the FED and FEC values. FED was calculated based on the concentrations of carbon monoxide, hydrogen cyanide and carbon dioxide measured during the tests and is shown in Figure 3. The data is uncorrected for the gas burner output, but this is consistent in both tests. FEC was calculated based on the concentrations of formaldehyde, acrolein, sulphur dioxide, hydrogen chloride and nitrous oxides and is shown in Figure 4. FED and FEC clearly demonstrate that the early phases of both tests, during which time the construction products were not yet involved in the fire, contribute the most to human toxicity. FED remained fairly static after about 10 mins. A slight increase in FEC occurred in the rock mineral fibre test after about 18 mins., but the increase and the absolute values were significantly lower than the peak values reached during flashover. This means that the contents of the room were the major contributors to both FED and FEC. Figure 3: FED comparison 7
Results Figure 4: FEC comparison 8
Conclusion Due to the identical set up of fire source and fire load in both tests, a number of conclusions can be drawn on the contributions of the building contents and the construction products used in these test build ups. Up until the point in time that the fire starts to decay (after 15 minutes), the building contents are the main contributor to the fire and cause the flash over situation. This results in a very similar heat release rate for the build up insulated with rock mineral wool and that with PIR. The contribution of the insulation materials is negligible since the plasterboard acts as a thermal barrier. This is expected for a typical room fire and demonstrates the need to look at the fire performance of whole build ups instead of construction products only. A sharp peak in the concentration of the toxic gases is reached when flash over occurs. At this stage, the concentrations of carbon monoxide and hydrogen cyanide briefly but clearly exceed the 30 minute LC 50 lethal concentration. During the decay phase the insulation material becomes exposed to the fire when the plasterboard starts to crack or fall down. This does not lead to new peak values in heat release rate for either of the build ups. Calculations show that the effluent from the armchair and other furniture contents are the major contributor to both FED and FEC values. The above mentioned observations are only valid for the tested build up and fire source, in particular: l for both tests, the ventilation conditions (opening of the doorway) were kept constant other ventilation conditions will have a major impact on the concentration of the toxic gases; l for both tests. an armchair was used as fire source a different fire source can have different concentrations and peak values of the measured effluent gases; and l depending on the fire load the insulation might contribute at a different point in time. The results demonstrate that efforts to reduce deaths from fires in dwellings should not focus solely on the performance of combustible construction products in general, and combustible insulation in particular. Efforts to improve the fire safety of dwelling contents and furnishings and the use of effective fire warning systems may have a much greater impact on fire safety. In the decay phase, carbon monoxide concentrations are very similar for both build ups. On the other hand, for rock mineral wool, the formaldehyde and acrolein concentrations are higher in the decay phase. For the PIR insulation, the hydrogen cyanide concentration is higher. However. all concentrations measured lie well below the 30 minute LC 50 lethal concentration in this decay phase. 9
Appendix A Room contents Both rooms were fully furnished. The contents of the rooms were: l curtain (fabric): installed width of 80 cm with a curtain rod (installed 15 cm from the wall and 5 cm from the ceiling); l armchair (timber and foam filling, 80 cm x 70 cm x 55 cm) with 2 pillows; l small table (timber, 55 cm x 55 cm x 50 cm) with a few magazines and a remote control; l TV bench (timber, 120 cm x 40 cm x 74 cm); l TV (19 inch); and l bookcase (timber, 40 cm x 28 cm x 202 cm) with 7 identical books. Contents of the tested rooms in relation to the burner Photos of the room content and the power socket 10
Appendix B Measurements Heat Release Rate The heat release was measured and recorded using the oxygen depletion method according to ISO 9705-1:2016 and the European standard EN 14390:2007. The sampling port was installed in the exhaust duct and oxygen and carbon dioxide were continuously measured which allowed for determination of the heat release rate. Concentration of Gaseous Effluents The fire effluent gases were measured with an FTIR spectrometer. The sampling and the analysis were done based on ISO 16405 (Room corner and open calorimeter Guidance on sampling and measurement of effluent gas production using FTIR technique) and ISO 19702 (Guidance for sampling and analysis of toxic gases and vapours in fire effluents using Fourier Transform Infrared (FTIR) spectroscopy). Gas sampling was installed in the duct work of the smoke exhaust system of the testing facility. This sampling position represents cooled and diluted fire effluents. ISO 16405 states that this sampling position is often preferred over sampling outflowing gases directly at the top of the doorway as matrix effects from the fire effluents are minimized by the dilution. At this position due to cooling and dilution effects all chemical reaction in the fire effluent would have ceased which would not be the case when sampling directly from the immediate vicinity of the fire where further chemical reaction, combustion and decomposition could take place. Using the known volume flow rate in the exhaust duct, the dilution effect has been reversed prior to calculating the Fractional Effective Dose (FED) and Fractional Effective Concentration (FEC). 11
Contact Details Customer Service For quotations, order placement and details of despatches please contact the Kingspan Insulation Customer Service Department on the numbers below: Tel: +44 (0) 1544 388 601 Fax: +44 (0) 1544 388 888 email: customerservice@kingspaninsulation.co.uk Literature & Samples Kingspan Insulation produces a comprehensive range of technical literature for specifi ers, contractors, stockists and end users. The literature contains clear user friendly advice on typical design; design considerations; thermal properties; sitework and product data. For copies please contact the Kingspan Insulation Marketing Department, or visit the Kingspan Insulation website, using the details below: Tel: +44 (0) 1544 387 384 Fax: +44 (0) 1544 387 484 email: literature@kingspaninsulation.co.uk www.kingspaninsulation.co.uk/literature Tapered Roofing For technical guidance, quotations, order placement and details of despatches please contact the Kingspan Insulation Tapered Roofi ng Department on the numbers below: Tel: +44 (0) 1544 387 383 Fax: +44 (0) 1544 387 483 email: tapered@kingspaninsulation.co.uk Technical Advice / Design Kingspan Insulation supports all of its products with a comprehensive Technical Advisory Service. Calculations can be carried out to provide U values, condensation / dew point risk, required insulation thicknesses etc U value calculations can also be carried out on the Kingspan Insulation U value Calculator, available for free online at www.uvalue-calculator.co.uk or downloaded as an App. The Kingspan Insulation Technical Service Department can also give general application advice and advice on design detailing and fi xing etc... Site surveys are also undertaken as appropriate. The Kingspan Insulation British Technical Service Department operates under a management system certifi ed to the BBA Scheme for Assessing the Competency of Persons to Undertake U value CERTIFICATE CS/1004-4 U Value Competency Scheme and Condensation Risk Calculations. Please contact the Kingspan Insulation Technical Service Department on the numbers below: Tel: +44 (0) 1544 387 382 Fax: +44 (0) 1544 387 482 email: technical@kingspaninsulation.co.uk General Enquiries For all other enquiries contact Kingspan Insulation on the numbers below: Tel: +44 (0) 1544 388 601 Fax: +44 (0) 1544 388 888 email: info@kingspaninsulation.co.uk Kingspan Insulation Ltd. reserves the right to amend product specifi cations without prior notice. Product thicknesses shown in this document should not be taken as being available ex stock and reference should be made to the current Kingspan Insulation price list or advice sought from Kingspan Insulation s Customer Service Department (see above left). The information, technical details and fi xing instructions etc. included in this literature are given in good faith and apply to uses described. Recommendations for use should be verifi ed for suitability and compliance with actual requirements, specifi cations and any applicable laws and regulations. For other applications or conditions of use, Kingspan Insulation offers a Technical Advisory Service (see above), the advice of which should be sought for uses of Kingspan Insulation products that are not specifi cally described herein. Please check that your copy of this literature is current by contacting the Kingspan Insulation Marketing Department (see left). Kingspan Insulation Ltd Pembridge, Leominster, Herefordshire HR6 9LA, UK www.kingspaninsulation.co.uk Kingspan and the Lion Device are Registered Trademarks of the Kingspan Group plc in the UK, Ireland and other countries. All rights reserved. Registered in England & Wales, No. 01882722. Registered Offi ce: Pembridge, Leominster, Herefordshire HR6 9LA UK. VAT GB428602456.