Understanding char loss and fall-off for fire-safe design of timber structures

Engineered timber is in high demand while the awareness of environmental pressures is growing. Worldwide research has studied the fire behaviour of timber but char fall-off remains one of the critical factors by delaying or preventing self-extinction. This article highlights our recent research findings and the need to further investigate the occurrence and effects of char fall-off to enable fire-safe design with engineered timber.

Building better

Engineered timber has the potential to play a key factor in reducing the carbon emissions of the built environment and construction industry.¹ Like any new technology, novel materials challenge existing norms: their design and delivery demand critical assessment of the intention of relevant design regulations, guidance and practice.²

Fire strategies associated with tall buildings require that the structure will retain stability indefinitely. This is because the environment must be safe for building occupants and firefighters for extended evacuation and intervention periods. For a combustible structure such as engineered timber, the designer must explicitly address the ability to resist compartment burnout and for the timber to self-extinguish.

The challenge associated with engineered timber is that if the timber burns, there is an explicit coupling between the structure and the compartment fire dynamics. This coupling means that it is no longer possible to define a rational period of fire resistance that is relatable to the burnout of the fuel load.² 

Burning of timber

Engineered timber structures may be exposed – to meet some aesthetic requirement – or encapsulated. Building codes around the world permit various uses and degrees of exposure depending on the perceived risks associated with a building. 

Encapsulation strategies are designed to protect the timber from burning and contributing to the compartment fire by preventing the transfer of heat to the timber and preventing pyrolysis from occurring. These systems must remain in place for the duration of the fire if these systems are to be used.

The burning of timber has been studied extensively and is reported in a wide body of literature.^3,4,5,6 The key characteristics are that timber is a charring material. Pyrolysis occurs around 200–350°C producing flammable gases and forming a rigid, carbon-rich char which is susceptible to oxidation. Authors often emphasise the insulating qualities and growth of the char layer. The char layer that forms on the surface of a heat-exposed piece of timber reduces the heat transfer to the virgin timber underneath. As a result, the production of pyrolysis gases slows down until flaming combustion is not sustained anymore – this is the principle of self-extinction.^3,4,7 Accounting for the char layer thickness is often perceived as a method to provide a sacrificial protective layer mitigating the combustible nature of timber and maintaining stability. 

Loss of the char layer

Char fall-off has been widely cited as one of the critical factors determining whether or not the timber structure will self-extinguish after burnout of the movable fuel load.^3,5 Solid timber elements will not sustain flaming combustion indefinitely without an external heat source. However, the burning behaviour of engineered timber does not necessarily mirror the same burning behaviour. The laminated nature of engineered timber means that the layers may detach from each other when the timber product is heated, burns or experiences increased loading. Consequently, engineered timber is more susceptible to the charred surface deteriorating or falling away from the underlying timber element, as the exemplary pictures in Figure 1 show. 

Although the occurrence of char loss has been widely documented in the literature, no complete theory of the underlying processes exists ⁸. Multiple mechanisms can cause the lamellae to separate, such as adhesive failure, or cohesive failure within the timber or glue layer. Since the exact mechanisms are not well identified, the term “char loss” (rather than “delamination”) is used to encompass any process which leads to the loss of charred material from the surface of a timber sample.

Consequences for timber buildings

Even if the exact mechanisms of char loss are not well understood, it is important that engineers can account for char loss in their design to ensure that the structure meets the requirement of sustaining stability to compartment burnout.

Loss of the char layer results in increased heat transfer to the underlying timber and the potential for sustained burning over long periods of time. The primary resulting hazard is loss of load-bearing capacity and integrity due to heating and pyrolysis of the timber structure. Charred timber is assumed to have negligible strength; the load-bearing capacity is reduced from temperatures as low as 60°C.⁵ Moreover, char fall-off increases the timber surface exposed to heating over time, increasing the overall fuel load and burning duration.

Experimental investigations/ Experimental study on char fall-off

We studied the thermal effect of char fall-off on the fire performance of engineered timber as part of a research project involving IMFSE, the University of Edinburgh and the University of Queensland. We quantified the thermal response of cross-laminated timber (CLT) by measuring the in-depth temperatures while the timber was heated and experienced char fall-off. 

Real-life structures will be subjected to loading when exposed to a fire. Nevertheless, most existing studies have used unloaded samples. With our experimental study, we wanted to investigate the impact of the worst-case loading condition compared to other loading scenarios.⁹

The bespoke test setup,^8,9 partly depicted in Figure 2, allowed us to test the fire behaviour of CLT under structural loads. The processes of thermal and mechanical degradation of timber and adhesives are linked.⁵ The eccentric axial load, applied through pin supports at both ends of the columns, induced a bending moment. This caused maximum stresses in the first glue line and lamella exposed to heating, which enhanced char fall-off. 

The energy received by the timber surface was calibrated via the distance between the panel and the timber surface. The samples were exposed to an incident heat flux of either 20kW/m², which is below the critical value of around 30kW/m² for self-extinction of timber,^3,7 or to 50kW/m². This allowed us to begin to understand the effect of different heating scenarios as would be expected in real fires.

 Half of the samples were tested as untreated timber columns. The other half of the samples were equipped with a layer of glass-fibre mat (623gsm), as shown in Figure 3. The glass-fibre holding any char pieces in place on the timber surface allowed us to assess the effect of char fall-off.

Glass fibre reinforced polymers (GFRP) have been shown to increase the structural performance of different materials, including laminated timber, at ambient conditions.^10,11 Glass-fibre mat embedded in epoxy polymer achieves this composite action, combining the strengths of both GFRP and timber at ambient conditions. 

Glass fibre is inert, which means it does not pyrolyse or oxidise. Therefore, we used the GFRP mat to mechanically prevent char fall-off while it was having minimal impact on the in-depth heat transfer. Since the polymer used for the GFRP decomposes at temperatures below the pyrolysis temperature of timber, metal staples were used to anchor the glass-fibre mat without relying on the adhesive when the timber reached elevated temperatures.

The samples’ surfaces were insulated except for the exact desired exposed area at the front surface where the incident heat flux was expected to be constant. Combined with the scale of the experiments, this ensured a high level of repeatability, one-dimensional heating and more accurate measurements of residual mass, char depths and in-depth temperatures. We placed 1.5mm K-type thermocouples at different depths below the heated CLT surface. The thermocouple orientation parallel to the exposed surface was expected to result in the smallest measurement error and least disturbance of the temperature field within the sample.^8,12 At the end of each test, the samples were immediately lightly misted with water to stop any flaming or smouldering combustion reactions. The cooling preserved the remaining timber and char, without any unwanted further destruction.

Results and discussion

The experimental study enabled a comparison between two extreme cases: CLT columns especially prone to char fall-off versus CLT columns that did not lose any char, both exposed to the same heating and loading conditions. The aim was to quantify the impact of char fall-off on the heating and burning of engineered timber. 

All CLT columns formed a char layer at both levels of heat exposure (20kW/m² and 50kW/m²). While all untreated CLT columns experienced char fall-off in every test, the layer of glass-fibre always successfully prevented char fall-off. The times from first heat exposure until each char fall-off event were analysed. The data analysis showed no clear trend regarding the measured glue line temperature and the occurrence of char fall-off. While the in-depth temperature evolution showed a high level of repeatability throughout the experimental study, the results do not support any direct correlation between char fall-off and the glass transition temperature of the adhesive or any other specific temperature.⁸ 

The samples that experienced char fall-off heated up faster in-depth than samples in which the protective char layer was held in place. Moreover, when char fall-off occurred, a larger fraction of the timber cross-section reached temperatures above which irreversible loss of structural capacity of the timber is expected. 

Figure 1 and Figure 4 show the remains of several CLT columns. Large amounts of the first lamella charred and fell off, exposing the underlying second lamella to the incident radiant heat flux. The surface of the second underlying lamella also shows charring. At the end of the test, the remaining char of the first lamella was fragile and only loosely attached to the second lamella. Moreover, flaming occurred behind parts of the remains of the first and second lamella. 

These figures also show clearly that char pieces fell off separately, not as a whole lamella at once. This observation proves that the term ‘delamination’ might not be the most accurate term to refer to this phenomenon.

Future research directions

Char fall-off is one of the key problems making the exploitation of the thermal insulation properties of char and the self-extinction of timber unreliable. Suitable research methods need to be developed to refine the understanding of char loss and char fall-off and its impact on timber structures in fire. 

The main research area that presents itself is the investigation of the controlling mechanisms. This could then enable us to learn how to predict and account for loss of the char layer in fire safety design of timber buildings. Moreover, measures to prevent char fall-off and reduce loss of the char layer can then be developed. The GFRP-timber-composite system used in this study is only one of many possible and potentially currently unthought-of methods. 

The key to making the methods to predict, account for and prevent char loss reliable is a sound fundamental understanding of the governing failure modes. Without this understanding, an immense number of tests would be needed to investigate the many different timber types, adhesives, manufacturing techniques, loading cases, fire scenarios and associated heating rates to just name a few factors that need to be taken into consideration in a hypothetical attempt to predict char fall-off without a thorough understanding of its controlling factors.

Combining current research efforts into a more systematic and meaningful approach based on a scientific investigation of the governing mechanisms has the potential to enable fire safe designs of engineered timber structures in the future.

Conclusions

Loss of the char layer through char fall-off can significantly enhance in-depth heat transfer within fire-exposed engineered timber. Therefore, char fall-off is very likely an important factor when assessing the structural capacity of engineered timber exposed to heating and fire. Moreover, char fall-off increased the rate of decomposition and burning of the timber structure itself. This affects the resulting fire dynamics and can result in extended burning durations, which means the fire could burn more intensely and for a longer time. As these phenomena are coupled, char fall-off calls to be accounted for in the design methodologies for timber buildings. However, the mechanisms behind char fall-off are still only poorly understood. No clear relation between glue line temperatures and char fall-off exists.

A thin layer of glass-fibre mat, tightly attached to the exposed surface, prevented any fall-off of char in the other half of the samples. There is an urgent need to identify ways to reliably prevent the occurrence of char fall-off due to the potential consequences regarding the loss of structural stability, compartmentation and the growth of the fire. Otherwise, char fall-off has to be accounted for in some quantitative way when designing with engineered timber. 

These results highlight the need for more detailed research on the occurrence and effects of char fall-off. To be able to design safe and sustainable timber buildings, it is clearly necessary to reliably prevent char fall-off or to account for it accurately. More targeted research focusing on the formation, oxidation and detachment of the char layer of heated and burning timber will be the next steps in the future.

Acknowledgements

Laura commenced working with WSP Australia in 2020. WSP prides itself on attracting and retaining the best talent who thrive on challenges and unconventional thinking in alignment with WSP’s Future Ready programme. WSP looks forward to continuing to support Laura’s career as she pursues holistic approaches to timber and fire safety.

This work was supported by the International Master of Science in Fire Safety Engineering (IMFSE) of Ghent University. This EU-funded programme provides students with a world-leading educational experience with advanced knowledge, skills and competences in fire safety engineering. This project was conducted as a joint effort between the University of Edinburgh and the University of Queensland.

For more information, go to www.wsp.com

References

1. Ramage, M. H. et al. (2017). The wood from the trees: The use of timber in construction. Renewable and Sustainable Energy Reviews. doi: 10.1016/j.rser.2016.09.107.
2. Hadden, R. M., and Law, A. (2020). We need to talk about timber. The Structural Engineer.
3. Bartlett, A. I., Hadden, R. M. and Bisby, L. A. (2019). A Review of Factors Affecting the Burning Behaviour of Wood for Application to Tall Timber Construction. Fire Technology. doi: 10.1007/s10694-018-0787-y.
4. Lowden, L., and Hull, T. (2013). Flammability behaviour of wood and a review of the methods for its reduction. Fire Sci. Rev., doi: 10.1186/2193-0414-2-4.
5. Wiesner, F. (2019). Structural behaviour of cross-laminated timber elements in fires. The University of Edinburgh.
6. Babrauskas, V. (2002). Ignition of wood: A review of the state of the art. J. Fire Prot. Eng. doi: 10.1177/10423910260620482.
7. Emberley, R., Inghelbrecht, A., Yu, Z. and Torero, J. L. (2017). Self-extinction of timber. Proc. Combust. Inst., vol. 36, no. 2, pp. 3055–3062. doi: 10.1016/j.proci.2016.07.077.
8. Schmidt, L. (2020). Experimental study on the effect of char fall-off on the heat transfer within loaded cross-laminated timber columns exposed to radiant heating. The University of Edinburgh & The International Master of Science in Fire Safety Engineering.
9. Chowdhury, P. (2019). Experimental study of char fall-off phenomenon in CLT under fire. The University of Queensland.
10. Miao C., D. Fernando, M. T. Heitzmann, and H. Bailleres, H. (2019). GFRP-to-timber bonded joints: Adhesive selection. Int. J. Adhes. Adhes., 2019, doi: 10.1016/j.ijadhadh.2019.05.007.
11. Fernando, D., Teng, J. G., Gattas, J. and Heitzmann, M. (2018). Hybrid fibre-reinforced polymer–timber thin-walled structural members. Adv. Struct. Eng. doi: 10.1177/1369433217739709.
12. Pope, I., Hidalgo, J.P., Torero, J.L. (2021). A correction method for thermal disturbances induced by thermocouples in a low-conductivity charring material. Fire Saf. J. 120. 103077. https://doi.org/10.1016/j.firesaf.2020.103077.