Cross-laminated timber, or CLT, is currently receiving attention for its potential use in tall building structures. Timber being a combustible material, one of the main challenges for the construction of these buildings is the potential fire risk that results from its use in the structure. For tall buildings, there is an increased severity of the consequences of structural failure as result of a fire. In order to obtain the desired level of safety, building codes typically require tall buildings to have a high fire resistance rating. The Eurocode recognises the risk associated with failure of the high rise structure and indicates that it is to be designed according to consequences class 3: failure of the structure due to fire is highly undesirable and the structure needs to be reliably safe and robust. The design of tall buildings in consequence class 3 requires a high degree of insight into their structural fire behaviour. An in-depth investigation is needed of risks associated with the fire behaviour, and the structural fire response. In this research the effect of using the combustible material CLT as the main bearing structure is investigated. As a combustible material, unprotected CLT can burn along with the fuel load present in a compartment. Irrespective of its fire resistance rating, it is uncertain whether the structure will be totally consumed in the event of a fire. This can result in collapse of the structure. It is important to understand whether the timber structure continues burning or whether a fire would decay by self-extinguishment, possibly in combination with active fire safety measures such as sprinkler activation or fire-brigade intervention. However, self-extinguishment currently is not part of fire safety considerations for the structural design of tall timber buildings. This master’s thesis aims to increase insight into the fire behaviour of unprotected CLT structures in a compartment burnout, conservatively assuming no active measures. The main research question of this work is: “Under what conditions is there a potential for self-extinguishment of cross-laminated timber?” A model of self-extinguishment of CLT was created which consists of various phases of a compartment burnout. Under the influence of an initial fire due to burning of room contents, the exposed CLT becomes involved in flaming combustion. Once the room contents have been largely consumed and the initial fire decays, the CLT contribution is expected to decrease as well, transforming from flaming to smouldering combustion. Finally, there will be a transition from smouldering to self-extinguishment. Two series of experiments were conducted to investigate this model and the conditions under which the transitions can take place. The first series of experiments investigated two conditions at which the CLT can transform from smouldering to self-extinguishment: heat flux on the CLT and the airflow over its surface. This was done by exposing CLT to various heat fluxes and airflows in a cone calorimeter. It was found that smouldering CLT self-extinguishes when the externally applied flux is below 5 to 6 kW/m2. The additional airflow was also found to be of influence. At a heat flux of 6 kW/m2, an airspeed of 0,5 m/s resulted in self-extinguishment, while an airspeed of 1,0 m/s resulted sustained smouldering. It can reasonably be assumed that at a heat flux below 6 kW/m2, the speed of the airflow should be limited to 0,5 m/s. The second series of experiments investigated all phases of the model of self-extinguishment. Small compartments with one, two, or three CLT walls were subjected to a propane fire with a decay phase. Delamination was found to be important in the fire behaviour of the CLT. Delamination occurred when the charring front reached the polyurethane adhesive, which then lost bonding. This resulted in fall-off and exposure of new layer of wood to the fire, which could interfere in the transitions from flaming to smouldering and from smouldering to self-extinguishment. In two experiments, fall-off occurred when temperatures in the compartment were still relatively high. The newly exposed wood contributed rapidly to the fire and flaming was sustained, or smouldering was transformed back to flaming. The CLT in these two experiments continued burning. In two other experiments, fall-off occurred when temperatures in the compartment were lower. As a result, flaming was not sustained. The CLT smouldered and the heat flux on its surface during was below 5 kW/m2; these compartments extinguished. However, relying on fall-off to occur when the compartment has cooled down might be risky due to its unpredictable nature. Alternatively, delamination and fall-off were prevented in a fifth experiment by applying a thicker top lamella. The charring front seized within the thickness of this lamella before it reached the PU adhesive. In all five compartments, the exposed CLT increased the heat release rate and total energy released, and extended the duration of a fire. These were further increased when fall-off resulted in prolonged flaming. Self-extinguishment is currently not part of fire safety considerations for timber high-rises. With further research, self-extinguishment might eventually contribute to a total fire safety concept for tall timber buildings, in addition to active measures. This research extends to translating new insights with regards to self-extinguishment to design and regulatory implications. A suggestion was made for a method to assess the potential for self-extinguishment of CLT structures. The method is based on the experiments and their specific conditions, e.g. with regards to the fire development, fuel load, ventilation conditions, CLT build-up, type of adhesive, wood species, and compartment configuration and scale. It would require further research and verification before the method could be applied in practice. The method consists of two steps. First, a thickness is determined, to be applied to the outer lamella of the CLT in order to prevent delamination. This thickness is based on the calculation of a finite charring depth for a parametric natural fire exposure, taking into account the ventilation conditions and the contribution of the CLT to the fuel load. Second, to allow the transformation to self-extinguishment, the compartment configuration should be such that the heat flux on the CLT surface is limited to 5 kW/m2 during smouldering. This can be achieved by estimating the temperatures of various surfaces and calculating the heat flux on the CLT with the most unfavourable configuration. With regards to building regulations and approval, self-extinguishment might contribute in demonstrating compliance of a timber structure with the Dutch building code. Self-extinguishment, as part of an increased insight in the structural fire behaviour, reduces the probability of collapse as a result of the fire. Despite severe consequences of collapse in a high-rise building, this reduction in probability might contribute in achieving a level of risk and safety as intended by the Dutch building code. Furthermore, an assessment of potential self-extinguishment might be considered in the design of timber buildings according to the Eurocode consequence class 3, which requires a high degree of insight into the structural fire behaviour. Self-extinguishment, as part of an overall fire safety concept, might become part of the risk analysis and the advanced calculations in order to demonstrate the tall timber structure is reliably safe and robust. Finally, new insight can affect the perception of the material and the way it is treated in building regulations and approval. Due to lack of precedents, limited experience, code limitations, and uncertainties regards the performance in fire, there seems to be reluctance to saying tall timber buildings are sufficiently safe. Timber is perceived as a material less suitable for high-rise buildings. More insight into self-extinguishment, its role in a total fire safety concept for tall timber buildings, and its implications on the structural design, could affect this. A more pragmatic approach might be adopted in which any material poses unique design challenges, but can be applied as long as it contributes to safe buildings. While the combustibility of the timber structure will remain a point of attention, its risk should be managed properly, instead of used to discard the material for high-rise buildings. Eventually, insight in the structural fire behaviour of timber structures, including potential self-extinguishment, could contribute to making timber available as a material for high-rise buildings.