*3.1. Bunsen Burner*

The temperature profiles and the development of char layers of the steel plates coated with intumescent coating formulations were recorded and compared. The data of time-temperature curves of the coatings are presented in Figure 3. Samples W, X, Y, and Z showed comparable temperature profiles after the test. During the first 10 min of fire, there was no difference in the temperature of all coating samples, and the temperature increased rapidly to 181, 170, 163 and 159 ◦C for samples W, X, Z, and Y, respectively. After 15 min of fire, the equilibrium temperatures were reached for all coatings and remained almost unchanged until 60 min of fire. The small-scale fire test results demonstrated that the equilibrium temperatures of curves W, X, Y, and Z were 173, 168, 155 and 160 ◦C, respectively.

**Figure 3.** The time-temperature curves of the coated steel plates with coating samples.

Moreover, Figure 4 exhibited that the thicknesses and expansion rates for samples Y and Z containing ES bio-filler were 39.50 mm-0.625 mm/min and 37.50 mm-0.59 mm/min, respectively. Sample Y had the best fire protection performance in terms of its equilibrium temperature and char formation compared to samples W, X, and Z. The growth of a multicellular char layer of sample Y was mainly attributed to the decarbonation of 3.5 wt.% of ES. It formed calcium oxide by releasing non-combustible carbon dioxide gas on heating, as follows:

$$\text{CaCO}\_3\text{ (s)} \rightarrow \text{CaO (s)} + \text{CO}\_2\text{ (g)}\tag{7}$$

**Figure 4.** The expansion rate and thickness of char layer of intumescent coating samples.

In addition, the expansion of the char layer can be initiated due to physical and chemical reactions contributed by an appropriate mixture of flame-retardant materials and binder, or the development of a cohesive structure during heating. This dense char layer could trap the degradation ingredients into the residue and result in a rounded swelling. This protecting layer declines the heat transfer from the heat source to the underlying steel in maintaining the integrity of the protected substrate against fire. The outcomes demonstrated that the coating comprising phosphate, nitrogen, ES, TiO2, Mg(OH)2, Al(OH)3 containing fire-retardant elements significantly contributed to a better fire protection performance, which resulted from the formation of a uniform and dense char layer. It was found that there was a correlation between the thickness of the char layer and the equilibrium temperature. This shows that the thickness of the char layer affected the fire protection performance of the coating.

#### *3.2. BS 476: Part 6*

The BS 476: Part 6 fire test found that all samples fulfilled the requirement, except the sample W ((*I*) = 22.3). The index and sub-index of the fire propagation test for all coating samples are presented in Table 2.


**Table 2.** The index and sub-index of BS 476: Part 6, fire propagation test.

The BS 476: Part 6 test results showed that the sub-index (*I*1:*I*2:*I*3) of coating samples W, X, Y and Z was (1.6:15:5.7), (0.2:5.1:1.5), (0.1:4.6:1.2) and (0:4.4:0.7), respectively. The index (*I*) results for the same coating samples were 22.3, 6.8, 5.9 and 5.1, respectively. It emphasized that the sub-index must be below 6 and the index of fire propagation must be below 12 for the coating samples to be certified as Class 0 materials. Among all coating samples, only sample W did not qualify as a Class 0 material since its index was 22.3 (*I*), which is out of the index of performance of this category (*I* > 12).

Evaluation of the fire propagation index for samples W, X, Y and Z revealed that samples Y ((*I*) = 5.9) and Z ((*I*) = 5.1) with 3.5 wt.% and 2.5 wt.% of ES, respectively, showed a great reduction in fire propagation index compared to sample W. It can be concluded that the incorporation of ES bio-filler into the coating formulation led to substantial inhibition of fire propagation, which could be contributed to the decarbonation of calcium carbonate at a high decomposition temperature [28].

In addition, sample X ((*I*) = 6.8) also exhibited a significant improvement in the reduction of fire propagation compared to sample W. This coating formulation showed an appropriate combination of TiO2/Al(OH)3/Mg(OH)2 flame retardant fillers with flame retardant ingredients led to a significant improvement in stopping the fire propagation behavior. This phenomenon is due to the main phosphorus element of ammonium polyphosphate (APP), which could easily respond with different oxides during a fire to produce ceramic-like solid materials (*X*-O-P species, *X* = Ti, B, Al, Mg, etc.). This develops a more cohesive and dense char structure [1,34,35]. The properties of the char structure are associated with fire protection performance of the sample [36,37].

#### *3.3. Cone Calorimeter Test*

The results of the cone calorimeter test using 50 kW/m<sup>2</sup> heat fluxes are shown in Table 3. The overall burning time of all intumescent coating samples was about 700 to over 900 s, and the TTI was 8–10 s.


**Table 3.** Data of the cone calorimeter test of samples.

According to cone measurements, the TTI values of samples W, X, Y, and Z were 9, 8, 10 and 10, respectively. The TTI of the high-density samples Y and Z, which contained ES bio-filler, had a longer time than those of the lower density sample X, demonstrating that the main factors were the density and decomposition temperature of the flame-retardant fillers [38]. In addition, the remaining mass of coating samples W, X, Y, and Z were 43.85%, 46.12%, 61.81% and 58.48%, respectively, after the test.

Therefore, samples Y and Z incorporated with ES were difficult to ignite and contribution to the TTI value and residual weight compared to samples W and X, due to its higher decomposition temperature. It is important to examine the profile of the HRR curve over time as it may reveal evidence on the varying thermal behavior of the heating process due to physical and chemical reactions of intumescent coatings.

Figure 5 displays the HRR versus time profiles of the coating samples after ignition. The burning behavior of the entire samples exhibited a single peak. Sample X showed the maximum peak of 111.86 kW/m<sup>2</sup> at 35 s, which was higher than other maximum peaks of samples W, Y and Z of 106.03, 91.00 and 99.98 kW/m2, respectively. The PHRR of sample Y was the lowest among all the samples due to its positive synergistic effect in reducing the heat release rate with the addition of 3.5 wt.% ES bio-filler into flame retardant ingredients and binder. The results show that the HRR value with the incorporation of bio-filler of samples Y and Z maintained at a level below 20 kW/m2 in the time range between 250–700 s ignition, while the HRR of samples W and X without addition of ES bio-filler decreased slowly and maintained at a level below 20 kW/m2 after at 600 s ignition.

**Figure 5.** Heat release rate (HRR) of the samples.

Figure 6 displays the curves of total heat released (THR) against the time of the intumescent coating samples. Samples W and X showed higher values of 22.4 and 21.6 MJ/m2, respectively, compared to samples B and D, which had lower values of 11.5 and 12.0 MJ/m2. The THR of samples W and X rose sharply and tended to follow a smooth curve after ignition. Samples Y and Z showed a very significant improvement in the reduction of the THR with the addition of the novel ES bio-filler. This result indicated that the heat release of samples Y and Z during combustion was very small and not enough to sustain combustion without external heat flux. The excellent flame retardancy properties of samples Y and Z were probably caused by the existence of the carboxylic group and calcium ions in the calcium carbonate and the carbon source, which promoted a dehydration reaction and decarboxylation reaction to release non-burning gases, such as H2O and CO2.

**Figure 6.** The curves of total heat released (THR) versus time of the samples.

The coating samples before and after the cone calorimeter test are shown in Figure 7a–d. Samples Y and Z, which comprise ES, had more effective char formation and expansion rate compared to samples W and X, due to appropriate combinations of flame retardant fillers (Y-ES/Al(OH)3/TiO2) and Z-ES/Al(OH)3/Mg(OH)2/TiO2). This could be attributed to the physical and chemical integration of the flame-retardant ingredients. The decomposition of Mg(OH)2 and Al(OH)3 flame-retardant fillers is described in the equations below:

$$\text{Mg(OH)}\_{2}\text{ (s)} \rightarrow \text{MgO (s)} + \text{H}\_{2}\text{O (g)}\tag{8}$$

$$2\text{Al(OH)}\_3\text{ (s)} \rightarrow \text{Al}\_2\text{O}\_3\text{ (s)} + 3\text{H}\_2\text{O}\text{ (g)}\tag{9}$$

**Figure 7.** The coating samples before (W, X, Y and Z) and after (**a**–**d**) the cone calorimeter test.

The properties of the Al(OH)3 flame-retardant filler displays strong reversibility of the dehydration reaction when exposed to heat, resulting in good fire resistance performance, since water released inside the particles recombine with the reactive surface of the freshly formed alumina [39]. However, the endothermic decomposition of the Mg(OH)2 filler would attribute to a gaseous water phase, which could enclose the flame by eliminating oxygen and dilute combustible gases by reducing the total heat released [40].

Sample Y revealed the highest rate growth and thickest char layer among the coating samples, as presented in Figure 7c. The development of the multicellular layer could have been originated by the release of non-combustible CO2 due to the decarbonation of ES bio-filler, which induces swelling by trapping the degradation products into the residue, as explained in Section 3.1.

The thermal degradation of ammonium polyphosphate can easily react with flame-retardant fillers to form a ceramic-like material, which increases the char formation by giving a dense and uniform char layer, which could insulate and protect the unprotected substrate in a fire [34,35].

### **4. Conclusions**

The thermal characteristics of four intumescent coating formulations have been studied in accordance with the BS 476 Part 6: Fire propagation test and ISO 5660-1 cone calorimeter standard test under atmospheric conditions with a piloted ignition. The incorporation of the ES bio-filler in the intumescent formulation led to a good thermal resistance and fire protection performance. It was found that all the parameters that characterize coating thermal resistance, such as TTI, HRR, and THR, decreased when 3.50 wt.% and 2.50 wt.% ES bio-filler was added to samples Y and Z. Hence, this study revealed that the addition of ES bio-filler strongly influenced the thermal properties and formation of the char layer of intumescent coatings. The coated samples X, Y and Z showed neither fire propagation nor afterglow combustion. Appropriate combinations of Al(OH)3/TiO2/ES in the coating formulation decreased the index value of fire propagation and HRR, whilst providing a thicker and more uniform char layer. The addition of renewable ES bio-filler showed significant enhancement in fire protection and the quality of the intumescent fire protective coatings, as well as being beneficial to the environment. In general, it can be determined that intumescent coatings display significant fire protection qualities in a practical and effective fire protective coating for steel, as shown by the findings of this study.

**Author Contributions:** Interpretation, J.H.B. and M.C.Y.; methodology, J.H.B.; validation, J.H.B., M.C.Y., M.K.Y. and L.H.S.; formal analysis, J.H.B.; investigation, J.H.B. and M.C.Y.; resources, J.H.B.; data collection, J.H.B.; writing—original draft preparation, J.H.B.; writing—review and editing, M.C.Y.; visualization, M.K.Y. and L.H.S.; supervision, M.C.Y.; project administration, M.C.Y.; funding acquisition, M.C.Y.

**Funding:** This project was funded by the University of Tunku Abdul Rahman under the UTARRF.

**Acknowledgments:** The authors would like to express their sincere gratitude to City University of Hong Kong for providing the laboratory services.

**Conflicts of Interest:** There is no conflict of interest declared by authors.

### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
