Intumescent Coatings Based on Tannins for Fire Protection

Accidents involving fire occur every day around the world, affecting thousands of people and causing economic losses. Some accidents are caused by steel structure failures, which experience a significant reduction in mechanical properties at temperatures of 400-550 ̊C. Therefore, fire protective coatings are required for steel structures and interest in the development of intumescent coatings has increased considerably. In this study, black wattle tannin was used as a carbon source in the formulation of intumescent coatings. Concentrations of 5% and 10% of tannin were incorporated into a novolac resin. The coating was applied on a steel plate and the thermal protection was evaluated by sample exposure to a flame for 30 min. The results showed that the tannin compound could be used as a carbon source for intumescent coatings. The temperature of the samples containing 10% of this compound was almost 300 ̊C lower compared to the uncoated steel plate.


Introduction
The protection of wood and steel structures against fire can save many lives in cases of disaster.The escape time within a confined space can be drastically reduced during a fire because steel structures experience a significant reduction in mechanical properties at temperatures higher than 450 °C 1,2 .Thus, the use of intumescent coatings is gaining interest as an approach to protecting these structures against fire [3][4][5] .
According to Alongi et al. 5 , the word "intumescent" refers to a material capable of expanding under certain conditions.Thus, intumescent coatings have the ability to expand when heated during a fire, creating a barrier to heat transfer and avoiding overheating of the substrate 6,7 .The composition of the intumescent system is described in detail in several reviews 5,8 , but in general it consists of three components: an inorganic acid source, an expanding or blowing agent and a carbon source 9 .The acid source is responsible for dehydrating the carbon source, which leads to the formation of char by the carbon source agent 9 .The expanding or blowing agent decomposes and releases non-flammable gases that are responsible for expanding the coating layer 9 .The carbon source typically has a substantial number of carbons and hydroxyl groups to enable the esterification reaction 7,9 .In this regard, renewable materials, such as tannin and lignin, can be useful as the carbon source in intumescent coatings.
Tannins are polyphenolic materials that are obtained from the bark, seeds and other parts of numerous species of trees 10,11 .Tannins are classified in two classes: condensed and hydrolysable 10,11 .Condensed tannins consist of chains of polyhydroxyflavan-3-ol units and are able to complex with metals, proteins and carbohydrates 12 .Hydrolysable tannins are polyesters based on gallic acid, with gallotannin and ellagitannin among the classes 11,13 .The potential of tannin usage is very wide, including medical 14,15 , pharmaceutical 16,17 and chemical 18,19 applications, as well as the materials industry 10,[20][21][22] .Our previous studies have reported the promising use of tannin as an antifouling pigment 23,24 and indicated that they can be highly effective in intumescent formulations because of the presence of high amounts of aromatic carbons and hydroxyl groups.
Recently, Alongi et al. 5 and Laoutid et al. 9 published interesting reviews and described recent research on intumescent coatings for many applications.Natural materials are also present in intumescent formulations [25][26][27][28] .Yew and Sulong 29 used chicken eggshell as a flame retardant filler in water-based intumescent coatings.Bodzay et al. 30 reported the use of mineral clay in formulations.Sources of carbon are also the subject of investigation.Ullah et al. 31 evaluated the effects of boric acid and melamine on intumescent coatings using expandable graphite as a carbon source.The presence of boric acid increased the amount of residual material on the remaining coating 31 .According to Gardelle et al. 32 , the presence of organoclay improved the mechanical properties of the char in intumescent coatings formulated with silicon.These authors also used expandable graphite as a carbon source, while Chou et al. 33 used pentaerythritol.To the best of our knowledge, there are no studies showing the use of tannins as a renewable and natural carbon source in intumescent coatings in the literature.
The aim of this work is to investigate the use of black wattle tannin as a carbon source in intumescent coatings formulated with epoxy novolac resin.Several formulations were compared a Laboratório de Materiais Poliméricos -LAPOL, Programa de Pós-Graduação em Engenharia de Minas, Metalúrgicae de Materiais -PPGE3M, Universidade Federal do Rio Grande do Sul, Av.Bento Gonçalves, 9500, 91501-970, Porto Alegre, RS, Brasil in a torch test.Complementary characterisation tests, such as X-ray diffraction (XRD) and thermogravimetric analyses (TGA), were also carried out to evaluate the oxidation of the steel substrate after the fire test and to characterise the thermal stability and determine the residual weight of the samples.These formulations allow the use of a compound of natural and renewable source that provides the formation of a carbonaceous layer rich in carbon, which acts as an insulating barrier and reduces the heat transfer from the fire to the underlying substrate.

Materials
Epoxy resin novolac DEN 431 solvent-free (Dow, USA) with an epoxy equivalent weight (EEW) of about 176 g/eq was used as a high solid binder in the intumescent coating formulation.Diethyltoluenediamine (Lonzacure (TM) 80) (Lonza, USA) with an amine hydrogen equivalent weight (AHEW) of 44.6 g/eq was used as an epoxy novolac hardener agent.The ratio of AHEW to EEW was 0.5.

Intumescent coating formulation
The compositions of epoxy novolac resin DEN 431 are given in Table 1.The tannin content was 5 wt% for the TAN5 sample and 10 wt% for the TAN10 sample; the BM samples were modified with 10 wt% boric acid and 10 wt% melamine; 5 wt% DOPO was added to P samples.First, epoxy resin novolac DEN 431 was heated to 100 °C and stirred constantly at 300 rpm using a mechanical stirrer model RW20 (IKA, Germany).Then, DOPO and triethanolamine were added and stirring continued at 110 °C for 30 min.After the reaction of DOPO and the epoxy resin, boric acid, melamine and black wattle tannin were added and the system was stirred for 30 min at 100 °C.

Sample preparation
Carbon steel substrates (AISI 1010) with dimensions of 100 mm × 100 mm × 0.9 mm were degreased with acetone and polished with #100 sandpaper.Then, the coating samples were applied using a spatula and a mask.Perforated aluminium sheets 1.5 mm thick and with holes 70 mm in diameter were used to establish a 1.5 mm coating.The steel substrates and the perforated aluminium sheets were sealed with adhesive tape.The samples were cured in an oven for 1 h at 140 °C, then the aluminium sheets were removed.The final thickness for the dry films (intumescent coatings) was 1.5 ± 0.1 mm, measured using a Byko-7500 thickness gauge (Byk-Gardner, Germany).

Fire resistance test and sample characterisation
The fire resistance test was conducted for each intumescent coating sample in triplicate.The increase in temperature after application of the flame was measured on the reverse of the metal substrate using a type K thermocouple (Thermomax, Brazil), as shown in Figure 1.
The temperature distribution on the reverse of the samples was carried out with a Fluke Ti400 thermography camera (Fluke, Canada) in the infrared spectral range.An acquisition module developed in the laboratory was used to perform collection of the thermocouple data.The flame was applied for 30 min using a Dremel model blowtorch, VersaFlame (Dremel, Germany).This butane torch reaches a maximum temperature of 1150 °C.
TGA of the samples was conducted using a thermogravimetric analyser, model TGA Q50 (TA Instruments, USA).The samples (about 10 mg) in the form of powder were placed in an open platinum pan.TGA experiments were carried out in inert atmosphere (N 2 ) within the temperature range 25 °C to 900 °C at a heating rate of 20 °C/min.
XRD analyses were performed to investigate the change in the steel surface (reverse) after application of the flame.The XRD patterns were obtained by analysing the reverse face of the substrate.The equipment used was the X'pert MPD (Philips, Netherlands) with CuK radiation and a scan rate of 0.05°/s.The test was performed on three samples: uncoated steel (blank) before the fire test, uncoated steel (blank) after the fire test and the steel plate coated with TAN10-BM-P intumescent coating after the fire test.

Fire resistance test
Initially, all formulations described in Table 1 were exposed to a blowtorch for 30 min.The temperature variation was recorded and is shown in Figure 2.
For all samples, as can be seen in Figure 2, the temperature of coatings stabilised after 3-6 min of fire testing.The blank sample reached 400 °C after 3 min of exposure to the flame, with the temperature increasing to values higher than 450 °C after 10 min (Figure 2).As previously mentioned, in this temperature range the structural properties of steel could be impaired 1,2 .Similar temperatures for the uncoated substrate (blank) were also found by Ullah et al. 31 .The pure epoxy resin novolac reached a temperature of approximately 330 °C after 5 min of the exposure to the flame.This decrease in the temperature compared to the uncoated steel is due to the high thermal stability of novolac resins (high crosslinking) 34 .According to Wang et al. 35 , novolac resin also improves flame resistance because it acts as a charring agent 36 .As can be seen in Figure 2, the TAN10-BM-P coating showed the largest reduction in the temperature of the steel (maximum of 147 °C).The influence of black wattle tannin in fire resistance is also clear.Comparing the TAN10-BM-P and TAN5-BM-P formulations, the presence of more tannin decreased the temperature on the reverse of the steel sample.The rate of heating on the reverse can be seen by the slope of initial times (up to 60 s) in Figure 2.
The slopes are as follows: 5.31 °C/s for the blank, 1.56 °C/s for epoxy resin, 1.21 °C/s for TAN5, 1.22 °C/s for TAN5-BM, 1.20 °C/s for TAN5-P, 0.76 °C/s for TAN5-BM-P, 0.92 °C/s for TAN10, 0.73 °C/s for TAN10-BM, 0.73 °C/s for TAN10-P and 0.69 °C/s for TAN10-BM-P.The slopes decreased sharply for TAN10-BM-P compared to the blank and the epoxy resin-coated sample.
Thermal images make it possible to observe the temperature distribution on the reverse surface and the heat affected zone of all samples (Figure 3).The highest temperatures were observed for the blank (Figure 3a) and the sample coated with epoxy resin (Figure 3b).The blank had the largest area affected by heat among all samples and had a maximum temperature of 557 °C after 30 min of exposure to flame.The epoxy resin sample showed slight substrate protection against the flame, with a reduction in temperature of 141 °C, corroborating the results of the fire resistance test.
The temperature distribution of the samples with 5% (dry mass) of black wattle tannin demonstrated an increase in fire resistance (Figure 4).It is possible to observe a considerable reduction in the temperature on the reverse surface due to the protection barrier produced by the intumescent effect.The interaction between tannin, melamine and boric acid promoted the formation of a char layer that insulated the metallic substrate from the flame.The system containing DOPO showed the lowest temperature; this can be explained by the in the gas phase, which suppresses the flame combustion radicals (H and OH) by a reaction with P-containing free radicals (P and PO) reducing the heat of the combustion 37 .
The temperature on the reverse of the samples for the formulations with 10% (dry mass) of black wattle tannin are shown in Figure 5.It can be observed that the effect of the temperature was reduced to the lowest level, mainly in the middle of the substrate.The TAN10-BM-P sample reached a temperature of around 110 °C after 30 min, confirming improved fire resistance.
The TGA emphasised the effect of each component on the intumescent system.The formulations with boric     acid and melamine (TAN-BM) presented a higher amount of residue compared to the other coatings, as displayed in Figure 6.The residue produced by the TAN10-BM sample was approximately 40% of initial weight, whereas for the epoxy resin sample the value was around 15%.This result indicates that the degradation of the tannin was incomplete and this behaviour provided a charred layer formation.This phenomenon was expected due to the presence of polyphenolic molecules, which are an effective carbon source 38 .
The first derivatives of the weight loss curves demonstrate that the intumescent samples (TAN-BM and TAN-BM-P) showed lower degradation kinetics than the others, which is certainly due to the formation of a protective layer.The degradation of the samples with the organophosphorus compound began at lower onset temperatures of 344 °C and 342 °C for the samples TAN5-BM-P and TAN10-BM-P, respectively, as can be seen in Table 2. Apparently, the phosphorus bonds for DOPO were chemically weaker than the energies of the resin bonds.In other words, there was lower activation energy for the decomposition of the material containing P 39 .The TAN10-BM sample showed a peak at 184 °C and 0.24 %/°C weight loss (Figure 6d), which indicates the dehydration of boric acid and consequently the formation of boron oxide (B 2 O 3 ), in line with Sevim et al. 40 .
The intumescent coatings containing DOPO (TAN-10BM-P) produced a lower amount of residue (32.02%), approximately 4.35% lower than TAN10-BM.This behaviour was expected since DOPO mainly acts in the gaseous phase in providing flame retardant action.Therefore, the organophosphorus must decompose to produce active radicals, conferring flame retardant performance previously described in the studies of Schartel et al. 41 .
The char layer structure of the samples after the burn test is shown in Figure 7.The substrate coated with epoxy resin was unprotected during the fire test due to the immediate degradation of the epoxy resin, which made contact between the fire and steel substrate possible, highlighted in Figure 7(a).The formulations with the tannin demonstrated an increase in residue and char formation promoted by its high aromaticity, which contributed to the charring effect and prevented the   of hydrogen and hydroxyl radicals by recombination of phosphorus radicals in the gas phase 42 .XRD demonstrated the presence of iron oxides on the steel without intumescent coating, as shown in Figure 8.This oxidation is a typical process when steel is exposed to high temperatures.In this case, the sample presented three iron oxides: magnetite (Fe 3 O 4 ), hematite (Fe 2 O 3 ) and wuestite (FeO).Magnetite and hematite are the products of oxidation between 300 °C and 600 °C 43 .When the oxidation process occurs above 570 °C, only wuestite is formed 44 .The presence of wuestite on the substrate indicates that the surface  has reached a temperature higher than 570 °C and this was confirmed by the thermal images.As the XRD pattern of the steel with intumescent coating TAN10-BM-P showed the same results as steel without heating, it is evident that this coating protected the substrate against fire damage.

Conclusion
The use of tannin in the intumescent coatings showed excellent results in terms of protection against fire.The thermal insulation barrier was provided by the synergistic effect of the constituents of the intumescent system.Moreover, the study demonstrated that using melamine and the organophosphorus compound as blowing agents modified the epoxy coating behaviour and changed it into an intumescent material.Furthermore, the best fire protection performance was found for the TAN10-BM-P sample, indicating that the addition of the tannin compound increased thermal insulation through the charring effect.XRD showed that the structure of the steel coated with TAN10-BM-P after the fire test was the same as the sample that had not been subjected to the fire test.Therefore, all components of the intumescent coatings had Materials Research a specific role and the combination of their characteristics allowed the development of an efficient fireproof protective coating.

Figure 1 .
Figure 1.Experimental scheme of the fire resistance test.

Figure 2 .
Figure 2. Temperature monitoring over time for different paint formulations.

Figure 3 .
Figure 3. Temperature distribution on coated sample after 30 min of blowtorch application: a) blank and b) epoxy resin.

Figure 6 .
Figure 6.Thermal degradation of samples: a) TG of the TAN5 set, b) DTG of the TAN5 set, c) TG of the TAN10 set and d) DTG of the TAN10 set.

Figure 8 .
Figure 8. X-ray diffraction patterns of the steel substrates after fire test.

Table 2 .
TG and DTG Data of intumescent systems under N2 atmosphere.