- Email: [email protected]
Thermal decomposition and fire response of non-halogenated polymer-based thermal coatings for concrete structures
SCT-21872; No of Pages 8 Surface & Coatings Technology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Thermal decomposition and fire response of non-halogenated polymer-based thermal coatings for concrete structures Yan Hao Ng a,b, Anil Suri a, Aravind Dasari a,⁎, Kang Hai Tan b,⁎ a b
School of Materials Science and Engineering (Blk N4.1), Nanyang Technological University, 50 Nanyang Avenue, Singapore 639789, Singapore School of Civil and Environmental Engineering (Blk N1), Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
a r t i c l e
i n f o
Article history: Received 31 August 2016 Revised 2 December 2016 Accepted in revised form 4 December 2016 Available online xxxx Keywords: Fire resistance Polymer coating Concrete
a b s t r a c t This work highlights the thermal decomposition and fire-protective behaviour of two polymer-based coatings for concrete. The coatings are entirely free of halogenated compounds. Although the combustion behaviour of the coatings is different from classical intumescence, their action is based on a condensed-phase mechanism, which creates a steep temperature gradient between the coating surface and the coating-concrete interface. When subjected to the ISO 834 heating curve, one of the coatings could prevent the temperature of the interface from rising above 345 °C even after 3 h (corresponding furnace temperature is 1114 °C). Heat transfer simulation corroborates the observed fire-protective behaviour and shows that it may have potential for use on structural members. Analysis of the concentration of several gases produced by the burning of the coatings shows that they do not pose an immediate danger to life and health. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Though concrete is non-combustible and has a relatively low value of thermal conductivity, use of thermal barriers or flame-retardant coatings on concrete maybe required in some special applications such as tunnel linings. This is because the compressive strength of concrete decreases with an increase in temperature (40% reduction at 500 °C and 96% reduction at 1000 °C for concrete with siliceous aggregates) [1,2]. Flame retardant (FR) coatings have gained attraction in the recent past due to increasingly stringent fire-resistance rating requirements based on hydrocarbon fire curve for structures. Otherwise, the building codes mandate the provision of sufficient concrete cover (and minimum cross-section dimensions) to ensure load-bearing capacity and integrity of concrete members during a fire [3]. In recent years, there is also a trend to use high strength concrete in construction. Exposing high-strength concrete with low permeability to fire permits pressure to build-up within its microstructure, and when the accumulating vapour pressure exceeds tensile strength of concrete, a sudden release of pressure results in explosive spalling [4]. Several methods such as embedding polypropylene fibres or providing supplementary reinforcement have been proposed to mitigate spalling. But for existing structures to meet stringent fire resistance requirement for alteration and addition works, thermal barrier is an effective way to reduce thermal gradient differences, and mitigate pore pressure and thermal stress spalling [5]. ⁎ Corresponding authors. E-mail addresses: [email protected] (A. Dasari), [email protected] (K.H. Tan).
Cementitious-based and geopolymer-based thermal barriers are some of the most widely explored passive fire protection solutions, particularly for tunnel linings [6–11]. However, the need for substantial increase in concrete cover (from 20 to 50 mm) will result in a significant increase in the weight of the member, which poses a major obstacle for these passive systems. The increase in thickness reduces tunnel gauge and affects the operational clearance of a tunnel [12]. This is where polymer-based coatings show great potential to replace traditional sprayed coatings. Halogenated compounds are widely regarded as effective FR additives, and 39% of the market is taken up by brominated FRs [13]. Although these systems provide an effective solution to reduce flammability of polymers, there are serious concerns of toxicity related to their application on-site and decomposition products in a fire. Besides, disposal of their residue poses environmental problems [13–15]. Intumescent systems, where materials swell and form a porous mass when exposed to fire or heat, are being increasingly considered for fire protection as an alternative to halogenated systems [16]. For these systems to intumesce and work effectively, three agents – an acid source, a carbonising agent, and a foaming agent, have to decompose systematically and in accordance with the matrix polymer. However, loss of cohesion of char structure and poor adhesion to substrate at high temperatures does not always guarantee the performance of intumescent systems [17]. Even the integrity of the swollen residue is a serious concern. This manuscript presents the performance of non-halogenated fireprotective coatings (developed in NTU laboratory) for structural applications. Although the coating does not exhibit classical intumescence behaviour, it relies on the condensed-phase mechanism to create a
http://dx.doi.org/10.1016/j.surfcoat.2016.12.015 0257-8972/© 2016 Elsevier B.V. All rights reserved.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
2
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
(two samples each from coatings A2 and B2) were used for the test and the average concentration values are presented in this paper.
huge temperature difference between the surface of coating and the base (interface between coating and substrate), thereby reducing heat transfer to the concrete substrate. Applying these coatings on existing concrete members improves their fire performance without adding excessive weight, and the effects of fire-induced spalling can be alleviated.
2.3. Direct flame torching The coatings were torched with a flame over a sustained period of 2 h to observe the behaviour and to determine their effectiveness as a thermal barrier. The temperature of the flame is calibrated to be 750 ± 50 °C before the test. Coatings with a thickness of ~2 mm were applied on two concrete blocks separately. The torching experiment was conducted in a chamber to ensure a constant atmospheric environment and heating conditions (flame temperature and flame length). Fig. 1 illustrates the experimental set-up and the locations of the thermocouples (represented as black dots). The coating was applied over a type-K thermocouple (thermocouple 1 in Fig. 1) that was embedded into a 50-mm thick concrete block. The thermocouple was flushed with the surface of the concrete and it recorded temperature of the concrete block. An additional type-K thermocouple (Thermocouple 2 in Fig. 1) was used to measure the temperature of the unexposed surface. The burner is placed such that the flame will be heating the area directly above the thermocouple. A laminar flame with minor flicking was produced using the burner, and the same area was heated consistently.
2. Experimental work 2.1. Coatings and their characterisation It is important to note that the emphasis of this manuscript is to explain the fire and thermo-oxidation performance of the coatings developed from an application viewpoint. Besides, considering the confidentiality of the materials/compositions of coatings (filed as a technology disclosure [18]), detailed information on the chemistry of these coatings will not be discussed here. Therefore, only a brief overview of the coatings is presented in Table 1. Thermogravimetric analysis (TGA) was used to investigate thermal stability of the coatings in air to simulate thermo-oxidative environment. The samples were analysed using TA Instruments Q500 from ambient temperature to 900 °C at a heating rate of 20 °C/min. Thermooxidative atmosphere represents the behaviour of the top most layer of the coating when exposed to an electric furnace in accordance to ISO 834 heating curve. Under direct flame exposure conditions, pyrolysis is expected to dominate within (or under) the exposed zone. Fourier transform infrared (FTIR) spectra of the coatings (fire exposed and unexposed surfaces) were obtained in attenuated total reflectance (ATR) mode using Perkin-Elmer Frontier Spectrometer (ATRFTIR). Three specimens for each coating were torched (exposed with a flame temperature calibrated to be 750 ± 50 °C) for a certain duration (10 min, 30 min, and 60 min) before the residues were removed for analysis. All spectra collected were in the range of 600 to 4000 cm−1 (at 0.5 cm−1 interval) under a resolution of 4 cm−1 and 32 scans per condition. Chemical changes on the surface exposed to flame can be studied using ATR-FTIR, which provides an indication of the coatings' decomposition over time.
2.4. Furnace test using ISO 834 heating curve A one-directional heat transfer experimental set-up is used to evaluate the fire performance of coatings on a concrete section according to the ISO 834 heating curve (Fig. 2a) [18]. For this purpose, a gypsum board fixture set-up was designed (see Fig. S1 and associated description in Supplementary Information) and placed at the opening of the electrical furnace. Thermal wool was also used to seal off any fine opening around this fixture to prevent heat loss to the surrounding area outside of the furnace, keeping the temperature within the furnace constant. Fig. 2b shows the schematic of a sectional view of the entire set-up. 3. Results and discussion This section describes the efforts to understand the complex thermal behaviour and fire performance of the coatings with the help of different thermo-analytical techniques.
2.2. Toxicity Toxicity of the gases released during decomposition of coatings is analysed using an air-tight test chamber constructed of an inert nonmetallic material (FESTEC NES 713 Toxic Chamber). Gases released will be retained in the chamber throughout the duration of the test, as any leakage will change their concentration and lead to inaccurate results. The burner within the set-up produces a flame height of 100 to 125 mm, and a flame temperature of 1000 ± 50 °C throughout the test. The entire specimen, weighing 0.10 ± 0.01 g, is engulfed in the flame during 1 min of burning. These conditions ensure complete combustion of the whole specimen before the mixing fan stirs the gases in the test chamber for 30 s, dispersing them evenly across the chamber. Gastec sampling pump (Model: GV-100S) connected to Gastec detection tubes are used to extract and analyse the concentration of a particular gas produced by the combustion process. A total of four samples
3.1. Thermo-oxidative behaviour From the mass loss curves in Fig. 3, it is evident that the addition of FR additives slightly reduces thermal stability of base matrices. The accurate drop in onset and maximum decomposition temperatures for all samples are listed in Table 2. It is worth nothing that, despite the reduction in thermal stability of coatings in the presence of FR additives, they show higher residual weight at 900 °C even in an oxidative environment. However, before striking off these coatings, it is important to appreciate the differences and the difficulties in correlating TGA data with practical tests like direct flame torching or ISO 834. Some of the
Table 1 Brief overview of the coatings discussed. Properties
Coating A
Base polymer/matrix Typical chemical structure
Polyurethane-based
Fire retardant additives
No
Coating A2
Coating B
Coating B2
Synthetic rubber-based
Yes
No
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Yes
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
3
Fig. 1. Experimental set-up for direct flame torching of coatings.
variations include sample size difference (mg in TGA versus bulk in ISO 834/torching), conditions of exposure (thermo-oxidation or fire), rate (20 °C/min in TGA versus direct flame exposure or increase in furnace temperature according to Fig. 2a in ISO 834 test) and even orientation (and confinement) of the samples. Besides, in a fire scenario, a one-dimensional temperature gradient can be expected to develop, in which only the uppermost surface of the coating is exposed to a high temperature, whereas the underlying portions may only be exposed to lower temperatures depending on the flame-retardant mechanism (condensed or gas phase). The difference will be exaggerated even further if filler orientation and participation in physical mechanisms like barrier formation are considered. Combustion response of montmorillonite (smectite clay) platelets in a polyamide 6 matrix is an excellent example
Fig. 3. (a) Mass loss of coating A, A2, B, and B2 in air environment; (b) Derivative mass loss rate of coating A, A2, B, and B2 in air environment.
for illustrating these differences and behaviour [19,20]. In these studies, despite the poor thermal stability of polyamide 6/clay nanocomposites (in TGA), heat release rates (tested in cone calorimeter) of nanocomposite were reduced by 50–70% compared to neat polymer, depending on the loading of clay and its distribution in the matrix. 3.1.1. Direct torching test Torching test conducted provides an indication of the coatings' effectiveness in creating a temperature difference between the heat source (fire) and the substrate. Fig. 4 presents the time-temperature profiles of coatings A2 and B2. As evident, the difference between the concrete temperature of both coatings is fairly constant for the first 30 min of torching; that is the rate of increase is relatively constant (1.23 °C/min for coating A2 and 1.99 °C/ Table 2 Summary of TGA results (average values) in air environment.
Fig. 2. (a) The relationship between average furnace temperature and time according to ISO 834 standard; and (b) Sectional view of the ISO 834 heating curve experimental setup.
Properties
Coating A
Coating A2
Coating B
Coating B2
T5% T50% Residual weight at 900 °C (%)
134.3 356.4 1.5
149.4 345.6 10.7
310.1 400.0 2.4
244.7 406.0 11.0
T5% - onset temperature, defined as the temperature at which 5% mass loss occurs; T50% - maximum decomposition temperature, defined as the temperature at which 50% mass loss occurs.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
4
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
Fig. 4. Time-temperature profiles for direct torching test.
min for coating B2). Subsequently, the rate of temperature increase dropped significantly in both cases. However, concrete temperature in coating B2 increased at a faster rate after 30 min and reached a temperature of ~117 °C at the end of 2 h (0.247 °C/min for coating B2 versus 0.09 °C/min for coating A2). Concrete temperature profile for coating A2 showed a plateau after 1 h of torching and reached only a temperature of ~ 66 °C at the end of 2 h. Temperature of the unexposed side (back face) of concrete only reached a maximum of ~40 °C at the end of 2 h. It should be noted that the temperature of the unexposed side might be influenced by heat dissipation to the surroundings (see Fig. 1) as normal-weight concrete has low thermal conductivity (λc = 1.6 W/(m·K) [21]). Coatings A and B had ignited and decomposed upon exposure to flame. The coatings were completely burned off after approximately 5 min, exposing the concrete surface (and therefore, no accurate timetemperature profiles could be obtained). Coatings A2 and B2 showed no such behaviour throughout the 2 h of torching. Charring in both coatings A2 and B2 was limited to uppermost layers, while most of the underlying coating seemed unaffected, an observation that appears to be supported by the FTIR results in Section 3.1.2. Surrounding areas not exposed to the flame seemed to be in a pristine condition suggesting that there was absolutely no flame spread.
3.1.2. FTIR results Results acquired using ATR-FTIR of the condensed phase (and residue) provide an indication of the chemical changes in the coatings across different torching durations. The spectra of coatings A2 and B2 presented in Fig. 5(a) and (b), respectively, indicate the outstanding behaviour of the coatings even after 2 h of direct flame exposure. There are three prominent peaks (1200 to 1000 cm−1, 1700 to 1500 cm−1, 3000 to 2800 cm−1) that are associated with pristine (not torched) coating A2 (Fig. 5a). These peaks remained unchanged after 10 min of torching. As the torching duration increases to 30 and 60 min, decomposition of CH3 compounds can be observed with a reduction in intensity of peak values between 3000 to 2800 cm− 1. The peaks between 1800 to 1300 cm−1 attributable to alkane C\\H stretch undergo changes over the duration of torching as the polymer chains undergo thermal fragmentation. Although the changes in carbonyl peak (1700 to 1600 cm−1) suggest thermal decomposition of urethane bond, the functional group is even present after 1 h of torching, as also corroborated by the presence of the C\\O\\C stretch at ~1100 cm−1, as well as the combination of N\\H bend and C\\N stretch at ~1250 cm−1 after 1 h.
Fig. 5. FTIR results for (a) coating A2; and (b) coating B2.
Spectra of coating B2 presented in Fig. 5b exhibits similar trends. A distinctive peak at 700 cm− 1, corresponding to alkene bend, can be seen across all 3 torching durations. The intensity of this peak at 60 min was slightly smaller compared to the intensity at 10 and 30 min. Two other prominent peaks observed between 1500 to 1400 cm−1 and 3000 to 2800 cm−1 were detected after 1 h of torching, suggesting minimal changes to the chemical composition of coating B2. Overall, the absence of any significant changes in the chemical composition of the coating as reflected in the IR spectra corroborates our view stated above that TGA data alone might not provide an insight into the fire-protection capacity of the coatings. 3.1.3. Toxicity Polymeric combustion products can affect the ability of people in the affected zone to evacuate or in some cases result in fatality [22]. In England, gas or smoke accounted for 40% of fire fatalities (cause of death was known) in 2014/15, a slight decrease from 47% in the previous year [23]. Thus, it is important to understand the concentration of the coatings' effluent in a fire. The average concentration of different gases for both coatings with FR additives (A2 and B2) and gas concentrations that are immediately dangerous to life or health (IDLH) are presented in Table 3. Interestingly, values of the concentration of CO2 detected after the combustion of coatings were similar to the production of CO2 from the methane burner alone (measured during calibration without a specimen in the furnace). FR additives could have interfered with the
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
5
Table 3 Concentration of different gases and IDLH values. Name
Formula
Coating A2 (ppm)
Coating B2 (ppm)
IDLH values [24] (ppm)
Carbon dioxide Carbon monoxide Nitrogen oxides Formaldehyde Sulphur dioxide Hydrogen sulphide Ammonia Phenol Hydrogen cyanide
CO2 CO NO + NO2 HCHO SO2 H2S
0 100 0.25 2.5 1.1 0.525
0 67.5 0.3 0.56 1.25 0.075
40,000 1200 20 (NO2) 20 100 100
NH3 C6H5OH HCNa
0 0 1.45
0 0 –
300 250 50
a HCN was not tested for coating B2 as there were no chemical substances in the coating that would contribute to the formation of its compounds.
decomposition process, resulting in the lack of detection of CO2. Future studies are required to precisely understand the roles of FR additives on the decomposition of polymer coatings and the resultant compounds released. As shown in Table 3, combustion of coatings A2 and B2 do not pose an immediate threat to a person's life or health as the concentration of different toxic gases emitted does not exceed the IDLH limit. In addition, the absence of CO2, NH3, and only traces of HCN gases may limit the danger posed to the occupants during a fire. It is well known that excessive inhalation of CO2 leads to disorientation or disturbed vision and hearing [25], and acute exposure to NH3 causes skin and eye irritation [26] that may impede an occupant's ability to evacuate.
3.2. Thermal performance under ISO 834 heating curve Coating A2 was exposed to ISO 834 heating curve for 3.5 h before the experiment was terminated. Four type-k thermocouples were placed across a vertical plane (as shown in Fig. S1b) to monitor the temperature surrounding coating A2 during the test. A schematic showing the positions of the thermocouples is given in Fig. 6a. The time-temperature profiles of the boundaries are shown in the Supporting Information section as Fig. S2. The purpose of obtaining boundary time-temperature profiles is to ascertain that heat transfer across the coating to the substrate, in the form of convection and radiation, is uniform. Fig. S2 clearly shows that the increase in temperature for all thermocouple was relatively similar throughout the test, suggesting that uniform temperature across the plane (1D heat transfer) was achieved. Fig. 6b, which plots the interface temperature as well as furnace temperature as a function of time, affirms that the gas temperature of the furnace closely matches the standard ISO 834 heating curve. Coating A2 performed exceptionally well as a thermal barrier throughout the test. Although the furnace gas temperature was ~1114 °C after 3 h, the interface temperature was only ~345 °C. This suggests that the coating was able to significantly increase the temperature difference and again contradicts the TGA data and confirms that many parameters influence the comparison at different scales. It should be noted that the residue of coating A2 remained adhering to the surface of the concrete after the set-up was removed approximately 17 h later. The concrete block remained intact except for minor cracks on the surface that were not detrimental to the integrity of the specimen.
3.3. Numerical simulation to assess the coating A performance on a structural slab As a prequel before conducting a large-scale structural fire test, numerical heat transfer simulation using ABAQUS software was carried out to assess the effectiveness of coating A2 on a concrete slab.
Fig. 6. (a) Schematic showing the positions of thermocouples; (b) Interface temperature profile during furnace test.
3.3.1. Input parameters and validation Heat transfer analysis in the simulation model accounts for conduction, convection and radiation. The input parameters for the simulation model were obtained from Eurocode, and the relevant clauses are summarised in Table 4. The model assumed a 1D heat transfer condition, with one surface of the slab exposed to heating in the form of convection and radiation based on ISO 834 time-temperature curve as illustrated in Fig. 7a. Heat conducted across the thickness of the slab was computed based on the specific heat capacity and thermal conductivity of concrete. The concrete slab model had a thickness of 200 mm (Fig. 7a) and the results obtained from the heat transfer analysis were validated using the R240 temperature profile in Fig. A.2 of Eurocode 2 (EC2) [2]. Fig. 7b presents a comparison of temperature profile obtained from the model with the temperature profile depicted in EC2. Results from the model demonstrate good agreement with the temperature profile in EC2. The upper limit of thermal conductivity of normal weight concrete was used in the model, which resulted in slight variations in temperature after 70 mm.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
6
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
Table 4 Input parameters from Eurocode. Parameter
Clause
Values
Specific heat capacity (J/kg·K)
3.3.2
Thermal conductivity (W/m·K) Emissivity of concrete Coefficient of heat transfer by convection (W/m2·K)
3.3.3
900 900 + (ϑ – 100) 1000 + (ϑ – 200)/2 1100 2–0.2451 (ϑ/100) + 0.0107 (ϑ/100)2
Eurocode/Reference
2.2(2) 3.2.1(2)
0.7 25
20 °C ≤ ϑ ≤ 100 °C 100 °C ≤ ϑ ≤ 200 °C 200 °C ≤ ϑ ≤ 400 °C 400 °C ≤ ϑ ≤ 1200 °C 20 °C ≤ ϑ ≤ 1200 °C
EN1992-1-2: 2004 [2]
EN1991-1-2: 2002 [27]
Specific heat capacity and thermal conductivity of concrete varies with temperature.
3.3.2. Heat transfer analysis Two numerical models with the same thickness (120 mm) were used as a case study to assess the effectiveness of coating A2. Heat transfer analysis for the first model (Model 1) was performed using convection and radiation based on ISO 834 heating curve for a duration of 3 h, depicting a concrete section that is not protected with the coating. Similar to the model used for validation, thermal conductivity and specific heat capacity were used to compute the heat conducted across the concrete slab. Heat transfer analysis for the second model (Model 2) was performed using the interface temperature profile obtained (Fig. 6b) from the experiment as an input for a duration of 3 h. Coating A2 has shielded the substrate from the heat source and created a huge temperature
difference between the furnace gas and the coating-concrete interface. Heat would be conducted across the thickness of the concrete block depending on its thermal conductivity and specific heat capacity. Hence, the interface temperature profile in Fig. 6b was used as an input and heat transfer analysis for the second model focused on heat conduction across the concrete section. Simulation results for Model 1 (unprotected concrete) and Model 2 (protected with coating A2) at the end of 3 h are illustrated in Fig. 8a and b, respectively. The temperature at the unexposed surface for Model 1 was computed to be 278.2 °C while the temperature for Model 2 was computed to be 90.2 °C, a difference of 188 °C. Time-temperature profile of the unexposed surface over the 3 h duration was presented in Fig. 9. Heat transfer has been drastically reduced due to the presence of the coating,
Fig. 7. (a) Illustration of heat transfer model used for validation; (b) Comparison of temperature profile given in EC2 (as Fig. A.2, Curve R240) with results from model.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
7
Fig. 8. Simulation results for (a) Model 1; and (b) Model 2.
which resulted in a slow increase in temperature across the concrete block. This created a remarkable reduction in maximum temperature at the end of 3 h and highlights the potential of coating A2 as a thermal barrier. 4. Summary Thermo-oxidative and fire performance of two novel fire-protective coatings for concrete were discussed in this paper. Both coatings are
non-halogenated and demonstrated remarkable performance when exposed to direct flame over a 2 h duration. The temperature of the concrete reached only 66 °C and 117 °C, respectively, for coating A2 and B2 at the end of 2 h flame exposure. Even the area exposed to fire on the sample was localised with no flame spread throughout the torching duration. Besides, analysis of the residues after direct flame torching of the coatings for different durations indicated the presence of polymer peaks. In the ISO 834 test (thermo-oxidative response), coating A2 performed well as a passive fire protection system on concrete structures. With a thickness of 10 mm, temperature at the concrete-coating interface was recorded to be only ~345 °C at the end of 3 h. More importantly, the concentration of several toxic gases emitted during the combustion of coatings was tested and found to pose no immediate threat to human life and health. Simulation results also validated the effectiveness of coating A2 on a normal grade concrete slab. Future research on the durability of the coatings (ageing, exposure to sun and rain) is necessary to ascertain the applicability and the performance of the coating.
Acknowledgement The authors gratefully acknowledge financial support by the Ministry of National Development - Singapore (Grant No. L2NICCFP1-20134). Appendix A. Supplementary data
Fig. 9. Comparison of time-temperature profiles obtained with Models 1 and 2 over a 3 h duration.
Details of the experimental set-up for the ISO834 standard fire test are available as electronic supplementary information. Supplementary data associated with this article can be found in the online version, at doi:10.1016/j.surfcoat.2016.12.015.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
8
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
References [1] V. Kodur, Properties of concrete at elevated temperatures, ISRN Civil Engineering 2014 (2014) 15. [2] European Comittee for Standarization, Eurocode 2: Design of Concrete StructuresPart 1–2: General Rules-Structural Fire DesignUnited Kingdom, London 2004. [3] European Comittee for Standarization, Eurocode 2: Design of Concrete Structures. Part 1–2: Structural Fire Design, CEN/TCUnited Kingdom, London 1993. [4] V.K.R. Kodur, Spalling in High Strength Concrete Exposed to Fire — Concerns, Causes, Critical Parameters and Cure, Structures Congress 2000Philadelphia, Pennsylvania, United States 2000. [5] G.A. Khoury, Passive fire protection of concrete structures, Proc. Inst. Civ. Eng. Struct. Build. 161 (2008) 135–145. [6] J.-H.J. Kim, Y. Mook Lim, J.P. Won, H.G. Park, Fire resistant behavior of newly developed bottom-ash-based cementitious coating applied concrete tunnel lining under RABT fire loading, Constr. Build. Mater. 24 (2010) 1984–1994. [7] V.R. Falikman, A.A. Shilin, Fire-protection coating of reinforced concrete lining in Lefortovsky Tunnel, in: P. Moyo (Ed.), 3rd International Conference on Concrete Repair, Rehabilitation and Retrofitting, ICCRRR-3, CRC Press 2012, Cape Town, South Africa 2005, pp. 265–268. [8] N.Ö. Bezgin, An experimental evaluation to determine the required thickness of passive fire protection layer for high strength concrete tunnel segments, Constr. Build. Mater. 95 (2015) 279–286. [9] L.F. Vilches, C. Leiva, J. Vale, C. Fernández-Pereira, Insulating capacity of fly ash pastes used for passive protection against fire, Cem. Concr. Compos. 27 (2005) 776–781. [10] K. Sakkas, D. Panias, P.P. Nomikos, A.I. Sofianos, Potassium based geopolymer for passive fire protection of concrete tunnels linings, Tunn. Undergr. Space Technol. 43 (2014) 148–156. [11] K. Sakkas, P. Nomikos, A. Sofianos, D. Panias, Inorganic polymeric materials for passive fire protection of underground constructions, Fire Mater. 37 (2013) 140–150. [12] Z.-G. Yan, H.-H. Zhu, J. Woody Ju, W.-Q. Ding, Full-scale fire tests of RC metro shield TBM tunnel linings, Constr. Build. Mater. 36 (2012) 484–494. [13] J. Murphy, Flame retardants: trends and new developments, Plast. Addit. Compd. 3 (2001) 16–20.
[14] M. Zhang, A. Buekens, X. Li, Brominated flame retardants and the formation of dioxins and furans in fires and combustion, J. Hazard. Mater. 304 (2016) 26–39. [15] R. Sühring, M. Freese, M. Schneider, S. Schubert, J.-D. Pohlmann, M. Alaee, H. Wolschke, R. Hanel, R. Ebinghaus, L. Marohn, Maternal transfer of emerging brominated and chlorinated flame retardants in European eels, Sci. Total Environ. 530– 531 (2015) 209–218. [16] J. Alongi, Z. Han, S. Bourbigot, Intumescence: tradition versus novelty. A comprehensive review, Prog. Polym. Sci. 51 (2015) 28–73. [17] M. Jimenez, S. Duquesne, S. Bourbigot, Characterization of the performance of an intumescent fire protective coating, Surf. Coat. Technol. 201 (2006) 979–987. [18] Non-halogenated fire-resistant coatings for concrete structures, Technology Disclosure, Ref.: TD/194/16, June 22, 2016, NTUitive Pte Ltd., Nanyang Technological University, Singapore. [19] A. Dasari, Z.-Z. Yu, Y.-W. Mai, G.P. Cai, H.H. Song, Roles of graphite oxide, clay and POSS during the combustion of polyamide 6, Polymer 50 (2009) 1577–1587. [20] I.S. Zope, A. Dasari, F. Guan, Z.-Z. Yu, Influence of metal ions on thermo-oxidative stability and combustion response of polyamide 6/clay nanocomposites, Polymer 92 (2016) 102–113. [21] Y. Wang, I. Burgess, F. Wald, M. Gillie, Performance-based Fire Engineering of Structures, CRC Press, 2012. [22] T.R. Hull, A.A. Stec, K. Lebek, D. Price, Factors affecting the combustion toxicity of polymeric materials, Polym. Degrad. Stab. 92 (2007) 2239–2246. [23] P. Gaught, J. Gallucci, G. Smalldridge, in: H.O. Statistics (Ed.), Fire Statistic England, 2014/15, 2016. [24] NIOSH, Pocket Guide to Chemical Hazards, 2008. [25] W.-R. Rupp, A. Thierauf, H. Nadjem, S. Vogt, Suicide by carbon dioxide, Forensic Sci. Int. 231 (2013) e30–e32. [26] J.E. Ryer-Powder, Health effects of ammonia, Plant/Oper. Prog. 10 (1991) 228–232. [27] European Comittee for Standarization, Eurocode 1: Actions on Structures Part 1–2: General Actions–Actions on Structures Exposed to FireUnited Kingdom, London 2002.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Thermal decomposition and fire response of non-halogenated polymer-based thermal coatings for concrete structures Yan Hao Ng a,b, Anil Suri a, Aravind Dasari a,⁎, Kang Hai Tan b,⁎ a b
School of Materials Science and Engineering (Blk N4.1), Nanyang Technological University, 50 Nanyang Avenue, Singapore 639789, Singapore School of Civil and Environmental Engineering (Blk N1), Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
a r t i c l e
i n f o
Article history: Received 31 August 2016 Revised 2 December 2016 Accepted in revised form 4 December 2016 Available online xxxx Keywords: Fire resistance Polymer coating Concrete
a b s t r a c t This work highlights the thermal decomposition and fire-protective behaviour of two polymer-based coatings for concrete. The coatings are entirely free of halogenated compounds. Although the combustion behaviour of the coatings is different from classical intumescence, their action is based on a condensed-phase mechanism, which creates a steep temperature gradient between the coating surface and the coating-concrete interface. When subjected to the ISO 834 heating curve, one of the coatings could prevent the temperature of the interface from rising above 345 °C even after 3 h (corresponding furnace temperature is 1114 °C). Heat transfer simulation corroborates the observed fire-protective behaviour and shows that it may have potential for use on structural members. Analysis of the concentration of several gases produced by the burning of the coatings shows that they do not pose an immediate danger to life and health. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Though concrete is non-combustible and has a relatively low value of thermal conductivity, use of thermal barriers or flame-retardant coatings on concrete maybe required in some special applications such as tunnel linings. This is because the compressive strength of concrete decreases with an increase in temperature (40% reduction at 500 °C and 96% reduction at 1000 °C for concrete with siliceous aggregates) [1,2]. Flame retardant (FR) coatings have gained attraction in the recent past due to increasingly stringent fire-resistance rating requirements based on hydrocarbon fire curve for structures. Otherwise, the building codes mandate the provision of sufficient concrete cover (and minimum cross-section dimensions) to ensure load-bearing capacity and integrity of concrete members during a fire [3]. In recent years, there is also a trend to use high strength concrete in construction. Exposing high-strength concrete with low permeability to fire permits pressure to build-up within its microstructure, and when the accumulating vapour pressure exceeds tensile strength of concrete, a sudden release of pressure results in explosive spalling [4]. Several methods such as embedding polypropylene fibres or providing supplementary reinforcement have been proposed to mitigate spalling. But for existing structures to meet stringent fire resistance requirement for alteration and addition works, thermal barrier is an effective way to reduce thermal gradient differences, and mitigate pore pressure and thermal stress spalling [5]. ⁎ Corresponding authors. E-mail addresses: [email protected] (A. Dasari), [email protected] (K.H. Tan).
Cementitious-based and geopolymer-based thermal barriers are some of the most widely explored passive fire protection solutions, particularly for tunnel linings [6–11]. However, the need for substantial increase in concrete cover (from 20 to 50 mm) will result in a significant increase in the weight of the member, which poses a major obstacle for these passive systems. The increase in thickness reduces tunnel gauge and affects the operational clearance of a tunnel [12]. This is where polymer-based coatings show great potential to replace traditional sprayed coatings. Halogenated compounds are widely regarded as effective FR additives, and 39% of the market is taken up by brominated FRs [13]. Although these systems provide an effective solution to reduce flammability of polymers, there are serious concerns of toxicity related to their application on-site and decomposition products in a fire. Besides, disposal of their residue poses environmental problems [13–15]. Intumescent systems, where materials swell and form a porous mass when exposed to fire or heat, are being increasingly considered for fire protection as an alternative to halogenated systems [16]. For these systems to intumesce and work effectively, three agents – an acid source, a carbonising agent, and a foaming agent, have to decompose systematically and in accordance with the matrix polymer. However, loss of cohesion of char structure and poor adhesion to substrate at high temperatures does not always guarantee the performance of intumescent systems [17]. Even the integrity of the swollen residue is a serious concern. This manuscript presents the performance of non-halogenated fireprotective coatings (developed in NTU laboratory) for structural applications. Although the coating does not exhibit classical intumescence behaviour, it relies on the condensed-phase mechanism to create a
http://dx.doi.org/10.1016/j.surfcoat.2016.12.015 0257-8972/© 2016 Elsevier B.V. All rights reserved.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
2
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
(two samples each from coatings A2 and B2) were used for the test and the average concentration values are presented in this paper.
huge temperature difference between the surface of coating and the base (interface between coating and substrate), thereby reducing heat transfer to the concrete substrate. Applying these coatings on existing concrete members improves their fire performance without adding excessive weight, and the effects of fire-induced spalling can be alleviated.
2.3. Direct flame torching The coatings were torched with a flame over a sustained period of 2 h to observe the behaviour and to determine their effectiveness as a thermal barrier. The temperature of the flame is calibrated to be 750 ± 50 °C before the test. Coatings with a thickness of ~2 mm were applied on two concrete blocks separately. The torching experiment was conducted in a chamber to ensure a constant atmospheric environment and heating conditions (flame temperature and flame length). Fig. 1 illustrates the experimental set-up and the locations of the thermocouples (represented as black dots). The coating was applied over a type-K thermocouple (thermocouple 1 in Fig. 1) that was embedded into a 50-mm thick concrete block. The thermocouple was flushed with the surface of the concrete and it recorded temperature of the concrete block. An additional type-K thermocouple (Thermocouple 2 in Fig. 1) was used to measure the temperature of the unexposed surface. The burner is placed such that the flame will be heating the area directly above the thermocouple. A laminar flame with minor flicking was produced using the burner, and the same area was heated consistently.
2. Experimental work 2.1. Coatings and their characterisation It is important to note that the emphasis of this manuscript is to explain the fire and thermo-oxidation performance of the coatings developed from an application viewpoint. Besides, considering the confidentiality of the materials/compositions of coatings (filed as a technology disclosure [18]), detailed information on the chemistry of these coatings will not be discussed here. Therefore, only a brief overview of the coatings is presented in Table 1. Thermogravimetric analysis (TGA) was used to investigate thermal stability of the coatings in air to simulate thermo-oxidative environment. The samples were analysed using TA Instruments Q500 from ambient temperature to 900 °C at a heating rate of 20 °C/min. Thermooxidative atmosphere represents the behaviour of the top most layer of the coating when exposed to an electric furnace in accordance to ISO 834 heating curve. Under direct flame exposure conditions, pyrolysis is expected to dominate within (or under) the exposed zone. Fourier transform infrared (FTIR) spectra of the coatings (fire exposed and unexposed surfaces) were obtained in attenuated total reflectance (ATR) mode using Perkin-Elmer Frontier Spectrometer (ATRFTIR). Three specimens for each coating were torched (exposed with a flame temperature calibrated to be 750 ± 50 °C) for a certain duration (10 min, 30 min, and 60 min) before the residues were removed for analysis. All spectra collected were in the range of 600 to 4000 cm−1 (at 0.5 cm−1 interval) under a resolution of 4 cm−1 and 32 scans per condition. Chemical changes on the surface exposed to flame can be studied using ATR-FTIR, which provides an indication of the coatings' decomposition over time.
2.4. Furnace test using ISO 834 heating curve A one-directional heat transfer experimental set-up is used to evaluate the fire performance of coatings on a concrete section according to the ISO 834 heating curve (Fig. 2a) [18]. For this purpose, a gypsum board fixture set-up was designed (see Fig. S1 and associated description in Supplementary Information) and placed at the opening of the electrical furnace. Thermal wool was also used to seal off any fine opening around this fixture to prevent heat loss to the surrounding area outside of the furnace, keeping the temperature within the furnace constant. Fig. 2b shows the schematic of a sectional view of the entire set-up. 3. Results and discussion This section describes the efforts to understand the complex thermal behaviour and fire performance of the coatings with the help of different thermo-analytical techniques.
2.2. Toxicity Toxicity of the gases released during decomposition of coatings is analysed using an air-tight test chamber constructed of an inert nonmetallic material (FESTEC NES 713 Toxic Chamber). Gases released will be retained in the chamber throughout the duration of the test, as any leakage will change their concentration and lead to inaccurate results. The burner within the set-up produces a flame height of 100 to 125 mm, and a flame temperature of 1000 ± 50 °C throughout the test. The entire specimen, weighing 0.10 ± 0.01 g, is engulfed in the flame during 1 min of burning. These conditions ensure complete combustion of the whole specimen before the mixing fan stirs the gases in the test chamber for 30 s, dispersing them evenly across the chamber. Gastec sampling pump (Model: GV-100S) connected to Gastec detection tubes are used to extract and analyse the concentration of a particular gas produced by the combustion process. A total of four samples
3.1. Thermo-oxidative behaviour From the mass loss curves in Fig. 3, it is evident that the addition of FR additives slightly reduces thermal stability of base matrices. The accurate drop in onset and maximum decomposition temperatures for all samples are listed in Table 2. It is worth nothing that, despite the reduction in thermal stability of coatings in the presence of FR additives, they show higher residual weight at 900 °C even in an oxidative environment. However, before striking off these coatings, it is important to appreciate the differences and the difficulties in correlating TGA data with practical tests like direct flame torching or ISO 834. Some of the
Table 1 Brief overview of the coatings discussed. Properties
Coating A
Base polymer/matrix Typical chemical structure
Polyurethane-based
Fire retardant additives
No
Coating A2
Coating B
Coating B2
Synthetic rubber-based
Yes
No
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Yes
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
3
Fig. 1. Experimental set-up for direct flame torching of coatings.
variations include sample size difference (mg in TGA versus bulk in ISO 834/torching), conditions of exposure (thermo-oxidation or fire), rate (20 °C/min in TGA versus direct flame exposure or increase in furnace temperature according to Fig. 2a in ISO 834 test) and even orientation (and confinement) of the samples. Besides, in a fire scenario, a one-dimensional temperature gradient can be expected to develop, in which only the uppermost surface of the coating is exposed to a high temperature, whereas the underlying portions may only be exposed to lower temperatures depending on the flame-retardant mechanism (condensed or gas phase). The difference will be exaggerated even further if filler orientation and participation in physical mechanisms like barrier formation are considered. Combustion response of montmorillonite (smectite clay) platelets in a polyamide 6 matrix is an excellent example
Fig. 3. (a) Mass loss of coating A, A2, B, and B2 in air environment; (b) Derivative mass loss rate of coating A, A2, B, and B2 in air environment.
for illustrating these differences and behaviour [19,20]. In these studies, despite the poor thermal stability of polyamide 6/clay nanocomposites (in TGA), heat release rates (tested in cone calorimeter) of nanocomposite were reduced by 50–70% compared to neat polymer, depending on the loading of clay and its distribution in the matrix. 3.1.1. Direct torching test Torching test conducted provides an indication of the coatings' effectiveness in creating a temperature difference between the heat source (fire) and the substrate. Fig. 4 presents the time-temperature profiles of coatings A2 and B2. As evident, the difference between the concrete temperature of both coatings is fairly constant for the first 30 min of torching; that is the rate of increase is relatively constant (1.23 °C/min for coating A2 and 1.99 °C/ Table 2 Summary of TGA results (average values) in air environment.
Fig. 2. (a) The relationship between average furnace temperature and time according to ISO 834 standard; and (b) Sectional view of the ISO 834 heating curve experimental setup.
Properties
Coating A
Coating A2
Coating B
Coating B2
T5% T50% Residual weight at 900 °C (%)
134.3 356.4 1.5
149.4 345.6 10.7
310.1 400.0 2.4
244.7 406.0 11.0
T5% - onset temperature, defined as the temperature at which 5% mass loss occurs; T50% - maximum decomposition temperature, defined as the temperature at which 50% mass loss occurs.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
4
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
Fig. 4. Time-temperature profiles for direct torching test.
min for coating B2). Subsequently, the rate of temperature increase dropped significantly in both cases. However, concrete temperature in coating B2 increased at a faster rate after 30 min and reached a temperature of ~117 °C at the end of 2 h (0.247 °C/min for coating B2 versus 0.09 °C/min for coating A2). Concrete temperature profile for coating A2 showed a plateau after 1 h of torching and reached only a temperature of ~ 66 °C at the end of 2 h. Temperature of the unexposed side (back face) of concrete only reached a maximum of ~40 °C at the end of 2 h. It should be noted that the temperature of the unexposed side might be influenced by heat dissipation to the surroundings (see Fig. 1) as normal-weight concrete has low thermal conductivity (λc = 1.6 W/(m·K) [21]). Coatings A and B had ignited and decomposed upon exposure to flame. The coatings were completely burned off after approximately 5 min, exposing the concrete surface (and therefore, no accurate timetemperature profiles could be obtained). Coatings A2 and B2 showed no such behaviour throughout the 2 h of torching. Charring in both coatings A2 and B2 was limited to uppermost layers, while most of the underlying coating seemed unaffected, an observation that appears to be supported by the FTIR results in Section 3.1.2. Surrounding areas not exposed to the flame seemed to be in a pristine condition suggesting that there was absolutely no flame spread.
3.1.2. FTIR results Results acquired using ATR-FTIR of the condensed phase (and residue) provide an indication of the chemical changes in the coatings across different torching durations. The spectra of coatings A2 and B2 presented in Fig. 5(a) and (b), respectively, indicate the outstanding behaviour of the coatings even after 2 h of direct flame exposure. There are three prominent peaks (1200 to 1000 cm−1, 1700 to 1500 cm−1, 3000 to 2800 cm−1) that are associated with pristine (not torched) coating A2 (Fig. 5a). These peaks remained unchanged after 10 min of torching. As the torching duration increases to 30 and 60 min, decomposition of CH3 compounds can be observed with a reduction in intensity of peak values between 3000 to 2800 cm− 1. The peaks between 1800 to 1300 cm−1 attributable to alkane C\\H stretch undergo changes over the duration of torching as the polymer chains undergo thermal fragmentation. Although the changes in carbonyl peak (1700 to 1600 cm−1) suggest thermal decomposition of urethane bond, the functional group is even present after 1 h of torching, as also corroborated by the presence of the C\\O\\C stretch at ~1100 cm−1, as well as the combination of N\\H bend and C\\N stretch at ~1250 cm−1 after 1 h.
Fig. 5. FTIR results for (a) coating A2; and (b) coating B2.
Spectra of coating B2 presented in Fig. 5b exhibits similar trends. A distinctive peak at 700 cm− 1, corresponding to alkene bend, can be seen across all 3 torching durations. The intensity of this peak at 60 min was slightly smaller compared to the intensity at 10 and 30 min. Two other prominent peaks observed between 1500 to 1400 cm−1 and 3000 to 2800 cm−1 were detected after 1 h of torching, suggesting minimal changes to the chemical composition of coating B2. Overall, the absence of any significant changes in the chemical composition of the coating as reflected in the IR spectra corroborates our view stated above that TGA data alone might not provide an insight into the fire-protection capacity of the coatings. 3.1.3. Toxicity Polymeric combustion products can affect the ability of people in the affected zone to evacuate or in some cases result in fatality [22]. In England, gas or smoke accounted for 40% of fire fatalities (cause of death was known) in 2014/15, a slight decrease from 47% in the previous year [23]. Thus, it is important to understand the concentration of the coatings' effluent in a fire. The average concentration of different gases for both coatings with FR additives (A2 and B2) and gas concentrations that are immediately dangerous to life or health (IDLH) are presented in Table 3. Interestingly, values of the concentration of CO2 detected after the combustion of coatings were similar to the production of CO2 from the methane burner alone (measured during calibration without a specimen in the furnace). FR additives could have interfered with the
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
5
Table 3 Concentration of different gases and IDLH values. Name
Formula
Coating A2 (ppm)
Coating B2 (ppm)
IDLH values [24] (ppm)
Carbon dioxide Carbon monoxide Nitrogen oxides Formaldehyde Sulphur dioxide Hydrogen sulphide Ammonia Phenol Hydrogen cyanide
CO2 CO NO + NO2 HCHO SO2 H2S
0 100 0.25 2.5 1.1 0.525
0 67.5 0.3 0.56 1.25 0.075
40,000 1200 20 (NO2) 20 100 100
NH3 C6H5OH HCNa
0 0 1.45
0 0 –
300 250 50
a HCN was not tested for coating B2 as there were no chemical substances in the coating that would contribute to the formation of its compounds.
decomposition process, resulting in the lack of detection of CO2. Future studies are required to precisely understand the roles of FR additives on the decomposition of polymer coatings and the resultant compounds released. As shown in Table 3, combustion of coatings A2 and B2 do not pose an immediate threat to a person's life or health as the concentration of different toxic gases emitted does not exceed the IDLH limit. In addition, the absence of CO2, NH3, and only traces of HCN gases may limit the danger posed to the occupants during a fire. It is well known that excessive inhalation of CO2 leads to disorientation or disturbed vision and hearing [25], and acute exposure to NH3 causes skin and eye irritation [26] that may impede an occupant's ability to evacuate.
3.2. Thermal performance under ISO 834 heating curve Coating A2 was exposed to ISO 834 heating curve for 3.5 h before the experiment was terminated. Four type-k thermocouples were placed across a vertical plane (as shown in Fig. S1b) to monitor the temperature surrounding coating A2 during the test. A schematic showing the positions of the thermocouples is given in Fig. 6a. The time-temperature profiles of the boundaries are shown in the Supporting Information section as Fig. S2. The purpose of obtaining boundary time-temperature profiles is to ascertain that heat transfer across the coating to the substrate, in the form of convection and radiation, is uniform. Fig. S2 clearly shows that the increase in temperature for all thermocouple was relatively similar throughout the test, suggesting that uniform temperature across the plane (1D heat transfer) was achieved. Fig. 6b, which plots the interface temperature as well as furnace temperature as a function of time, affirms that the gas temperature of the furnace closely matches the standard ISO 834 heating curve. Coating A2 performed exceptionally well as a thermal barrier throughout the test. Although the furnace gas temperature was ~1114 °C after 3 h, the interface temperature was only ~345 °C. This suggests that the coating was able to significantly increase the temperature difference and again contradicts the TGA data and confirms that many parameters influence the comparison at different scales. It should be noted that the residue of coating A2 remained adhering to the surface of the concrete after the set-up was removed approximately 17 h later. The concrete block remained intact except for minor cracks on the surface that were not detrimental to the integrity of the specimen.
3.3. Numerical simulation to assess the coating A performance on a structural slab As a prequel before conducting a large-scale structural fire test, numerical heat transfer simulation using ABAQUS software was carried out to assess the effectiveness of coating A2 on a concrete slab.
Fig. 6. (a) Schematic showing the positions of thermocouples; (b) Interface temperature profile during furnace test.
3.3.1. Input parameters and validation Heat transfer analysis in the simulation model accounts for conduction, convection and radiation. The input parameters for the simulation model were obtained from Eurocode, and the relevant clauses are summarised in Table 4. The model assumed a 1D heat transfer condition, with one surface of the slab exposed to heating in the form of convection and radiation based on ISO 834 time-temperature curve as illustrated in Fig. 7a. Heat conducted across the thickness of the slab was computed based on the specific heat capacity and thermal conductivity of concrete. The concrete slab model had a thickness of 200 mm (Fig. 7a) and the results obtained from the heat transfer analysis were validated using the R240 temperature profile in Fig. A.2 of Eurocode 2 (EC2) [2]. Fig. 7b presents a comparison of temperature profile obtained from the model with the temperature profile depicted in EC2. Results from the model demonstrate good agreement with the temperature profile in EC2. The upper limit of thermal conductivity of normal weight concrete was used in the model, which resulted in slight variations in temperature after 70 mm.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
6
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
Table 4 Input parameters from Eurocode. Parameter
Clause
Values
Specific heat capacity (J/kg·K)
3.3.2
Thermal conductivity (W/m·K) Emissivity of concrete Coefficient of heat transfer by convection (W/m2·K)
3.3.3
900 900 + (ϑ – 100) 1000 + (ϑ – 200)/2 1100 2–0.2451 (ϑ/100) + 0.0107 (ϑ/100)2
Eurocode/Reference
2.2(2) 3.2.1(2)
0.7 25
20 °C ≤ ϑ ≤ 100 °C 100 °C ≤ ϑ ≤ 200 °C 200 °C ≤ ϑ ≤ 400 °C 400 °C ≤ ϑ ≤ 1200 °C 20 °C ≤ ϑ ≤ 1200 °C
EN1992-1-2: 2004 [2]
EN1991-1-2: 2002 [27]
Specific heat capacity and thermal conductivity of concrete varies with temperature.
3.3.2. Heat transfer analysis Two numerical models with the same thickness (120 mm) were used as a case study to assess the effectiveness of coating A2. Heat transfer analysis for the first model (Model 1) was performed using convection and radiation based on ISO 834 heating curve for a duration of 3 h, depicting a concrete section that is not protected with the coating. Similar to the model used for validation, thermal conductivity and specific heat capacity were used to compute the heat conducted across the concrete slab. Heat transfer analysis for the second model (Model 2) was performed using the interface temperature profile obtained (Fig. 6b) from the experiment as an input for a duration of 3 h. Coating A2 has shielded the substrate from the heat source and created a huge temperature
difference between the furnace gas and the coating-concrete interface. Heat would be conducted across the thickness of the concrete block depending on its thermal conductivity and specific heat capacity. Hence, the interface temperature profile in Fig. 6b was used as an input and heat transfer analysis for the second model focused on heat conduction across the concrete section. Simulation results for Model 1 (unprotected concrete) and Model 2 (protected with coating A2) at the end of 3 h are illustrated in Fig. 8a and b, respectively. The temperature at the unexposed surface for Model 1 was computed to be 278.2 °C while the temperature for Model 2 was computed to be 90.2 °C, a difference of 188 °C. Time-temperature profile of the unexposed surface over the 3 h duration was presented in Fig. 9. Heat transfer has been drastically reduced due to the presence of the coating,
Fig. 7. (a) Illustration of heat transfer model used for validation; (b) Comparison of temperature profile given in EC2 (as Fig. A.2, Curve R240) with results from model.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
7
Fig. 8. Simulation results for (a) Model 1; and (b) Model 2.
which resulted in a slow increase in temperature across the concrete block. This created a remarkable reduction in maximum temperature at the end of 3 h and highlights the potential of coating A2 as a thermal barrier. 4. Summary Thermo-oxidative and fire performance of two novel fire-protective coatings for concrete were discussed in this paper. Both coatings are
non-halogenated and demonstrated remarkable performance when exposed to direct flame over a 2 h duration. The temperature of the concrete reached only 66 °C and 117 °C, respectively, for coating A2 and B2 at the end of 2 h flame exposure. Even the area exposed to fire on the sample was localised with no flame spread throughout the torching duration. Besides, analysis of the residues after direct flame torching of the coatings for different durations indicated the presence of polymer peaks. In the ISO 834 test (thermo-oxidative response), coating A2 performed well as a passive fire protection system on concrete structures. With a thickness of 10 mm, temperature at the concrete-coating interface was recorded to be only ~345 °C at the end of 3 h. More importantly, the concentration of several toxic gases emitted during the combustion of coatings was tested and found to pose no immediate threat to human life and health. Simulation results also validated the effectiveness of coating A2 on a normal grade concrete slab. Future research on the durability of the coatings (ageing, exposure to sun and rain) is necessary to ascertain the applicability and the performance of the coating.
Acknowledgement The authors gratefully acknowledge financial support by the Ministry of National Development - Singapore (Grant No. L2NICCFP1-20134). Appendix A. Supplementary data
Fig. 9. Comparison of time-temperature profiles obtained with Models 1 and 2 over a 3 h duration.
Details of the experimental set-up for the ISO834 standard fire test are available as electronic supplementary information. Supplementary data associated with this article can be found in the online version, at doi:10.1016/j.surfcoat.2016.12.015.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015
8
Y.H. Ng et al. / Surface & Coatings Technology xxx (2016) xxx–xxx
References [1] V. Kodur, Properties of concrete at elevated temperatures, ISRN Civil Engineering 2014 (2014) 15. [2] European Comittee for Standarization, Eurocode 2: Design of Concrete StructuresPart 1–2: General Rules-Structural Fire DesignUnited Kingdom, London 2004. [3] European Comittee for Standarization, Eurocode 2: Design of Concrete Structures. Part 1–2: Structural Fire Design, CEN/TCUnited Kingdom, London 1993. [4] V.K.R. Kodur, Spalling in High Strength Concrete Exposed to Fire — Concerns, Causes, Critical Parameters and Cure, Structures Congress 2000Philadelphia, Pennsylvania, United States 2000. [5] G.A. Khoury, Passive fire protection of concrete structures, Proc. Inst. Civ. Eng. Struct. Build. 161 (2008) 135–145. [6] J.-H.J. Kim, Y. Mook Lim, J.P. Won, H.G. Park, Fire resistant behavior of newly developed bottom-ash-based cementitious coating applied concrete tunnel lining under RABT fire loading, Constr. Build. Mater. 24 (2010) 1984–1994. [7] V.R. Falikman, A.A. Shilin, Fire-protection coating of reinforced concrete lining in Lefortovsky Tunnel, in: P. Moyo (Ed.), 3rd International Conference on Concrete Repair, Rehabilitation and Retrofitting, ICCRRR-3, CRC Press 2012, Cape Town, South Africa 2005, pp. 265–268. [8] N.Ö. Bezgin, An experimental evaluation to determine the required thickness of passive fire protection layer for high strength concrete tunnel segments, Constr. Build. Mater. 95 (2015) 279–286. [9] L.F. Vilches, C. Leiva, J. Vale, C. Fernández-Pereira, Insulating capacity of fly ash pastes used for passive protection against fire, Cem. Concr. Compos. 27 (2005) 776–781. [10] K. Sakkas, D. Panias, P.P. Nomikos, A.I. Sofianos, Potassium based geopolymer for passive fire protection of concrete tunnels linings, Tunn. Undergr. Space Technol. 43 (2014) 148–156. [11] K. Sakkas, P. Nomikos, A. Sofianos, D. Panias, Inorganic polymeric materials for passive fire protection of underground constructions, Fire Mater. 37 (2013) 140–150. [12] Z.-G. Yan, H.-H. Zhu, J. Woody Ju, W.-Q. Ding, Full-scale fire tests of RC metro shield TBM tunnel linings, Constr. Build. Mater. 36 (2012) 484–494. [13] J. Murphy, Flame retardants: trends and new developments, Plast. Addit. Compd. 3 (2001) 16–20.
[14] M. Zhang, A. Buekens, X. Li, Brominated flame retardants and the formation of dioxins and furans in fires and combustion, J. Hazard. Mater. 304 (2016) 26–39. [15] R. Sühring, M. Freese, M. Schneider, S. Schubert, J.-D. Pohlmann, M. Alaee, H. Wolschke, R. Hanel, R. Ebinghaus, L. Marohn, Maternal transfer of emerging brominated and chlorinated flame retardants in European eels, Sci. Total Environ. 530– 531 (2015) 209–218. [16] J. Alongi, Z. Han, S. Bourbigot, Intumescence: tradition versus novelty. A comprehensive review, Prog. Polym. Sci. 51 (2015) 28–73. [17] M. Jimenez, S. Duquesne, S. Bourbigot, Characterization of the performance of an intumescent fire protective coating, Surf. Coat. Technol. 201 (2006) 979–987. [18] Non-halogenated fire-resistant coatings for concrete structures, Technology Disclosure, Ref.: TD/194/16, June 22, 2016, NTUitive Pte Ltd., Nanyang Technological University, Singapore. [19] A. Dasari, Z.-Z. Yu, Y.-W. Mai, G.P. Cai, H.H. Song, Roles of graphite oxide, clay and POSS during the combustion of polyamide 6, Polymer 50 (2009) 1577–1587. [20] I.S. Zope, A. Dasari, F. Guan, Z.-Z. Yu, Influence of metal ions on thermo-oxidative stability and combustion response of polyamide 6/clay nanocomposites, Polymer 92 (2016) 102–113. [21] Y. Wang, I. Burgess, F. Wald, M. Gillie, Performance-based Fire Engineering of Structures, CRC Press, 2012. [22] T.R. Hull, A.A. Stec, K. Lebek, D. Price, Factors affecting the combustion toxicity of polymeric materials, Polym. Degrad. Stab. 92 (2007) 2239–2246. [23] P. Gaught, J. Gallucci, G. Smalldridge, in: H.O. Statistics (Ed.), Fire Statistic England, 2014/15, 2016. [24] NIOSH, Pocket Guide to Chemical Hazards, 2008. [25] W.-R. Rupp, A. Thierauf, H. Nadjem, S. Vogt, Suicide by carbon dioxide, Forensic Sci. Int. 231 (2013) e30–e32. [26] J.E. Ryer-Powder, Health effects of ammonia, Plant/Oper. Prog. 10 (1991) 228–232. [27] European Comittee for Standarization, Eurocode 1: Actions on Structures Part 1–2: General Actions–Actions on Structures Exposed to FireUnited Kingdom, London 2002.
Please cite this article as: Y.H. Ng, et al., Surf. Coat. Technol. (2016), http://dx.doi.org/10.1016/j.surfcoat.2016.12.015