The main purpose of this thesis is to find a future direction for the engine project for the FWM of the Atalanta project. Based on previous work done by Arjan Meskers [60], it was indicated that hydrogen peroxide is a good candidate as energy source for small scale engines because of its ease of implementation and relatively high energy content. Using hydrogen peroxide as a starting point, further research has to be done to explore different possibilities for the FWM engine. The first step is done in Chapter 1: an exploration of other projects on a similar scale. By learning from these projects that are described in literature, it is observed that certain directions are certainly not suitable for the FWM engine. For example, the bladed turbine has not the potential to become the FWM engine since the performance is too low and the requirements to make the system work are too substantial. Also, there is a clear indication that engines with a traditional cylinder piston assembly are dominated by leakage effects and therefore are also not successful on small scale. With this information a subset of the possibilities for the FWM engine is assembled in the form of concepts. These concepts are selected on their potential on small scale, their potential to be realized in a relatively short time span and their diversity. A candidate that does have potential is the small scale heat engine, represented by Concept 3. The performance is too low of similar projects found in literature, but they are not optimized for the FWM situation. Also, an indication in literature is found that the power density of these types of engines has good scaling behavior. Another potential candidate for the FWM is formed by applying a solution for the leakage problems of the small scale piston cylinder assembly. In Concept 1 flexible material is used between a piston and cylinder such that no fluid can leak through the gap. The blade-less turbine, also called a Tesla turbine, is used in Concept 2. It is described in literature to have a good theoretical potential on small scale, although not much experimental projects are found. The next step is to indicate what determines the performance characteristics for these small scale engines and how it relates to the requirements for the FWM engine. This is done in Chapter 2. The very basics for the thermodynamic cycle theory is the Carnot cycle, which closely resembles the operation principle of Concept 3. This is used as a starting point for approximating the performance of the concepts. The extension of the Carnot cycle is the Curzon Ahlborn model, which shows that the performance is completely determined by a certain potential and the utilization of that potential. The potential determines the magnitude of the incoming energy flow. The utilization determines how much of that incoming energy is converted into useful mechanical work. The nondimensionalization of the Curzon Ahlborn model shows that the utilization of the engine does not depend on any absolute scale, but merely on the ratio’s between certain engine parameters. These findings are tested on a more complicated model of the same system. By introducing the energy balance of the working fluid into the model, a time response of its temperature is obtained. This model shows that similar characteristics can be expected regarding the potential and utilization. This information is used to formulate a search method for finding the optimal configuration of the engine model to ensure maximum utilization for a given potential. By making an estimate of the engine potential based on measurements found in literature, this search method is used to give an indication of the performance characteristics as function of the scale of the engine. The study of Curzon Ahlborn type models is mainly focused on engines that have a compression and expansion step in the cycle and engines that use heat conduction as main energy transfer mode. Also all the models in the beginning of Chapter 2 are based on the assumption that work is extracted by a pressure force. Concept 2, the Tesla turbine, uses a different method of work extraction and has no compression step. Therefore, a separate model is presented at the end of Chapter 2 to characterize this concept. It is observed how the power output can be improved and how the efficiency is influenced accordantly. Measurements are found in literature from Tesla turbines at the exact scale that is opted for the FWM engine. These measurements are done with very small pressure differences, because the tests are done with a different application in mind. By combining the measurement results and the information from the simple model it is concluded that Concept 2 has no potential as FWM engine. One of the selection criteria for the engine concepts was that they all have different operating principles. Consequently, they all have different types of energy inflows. The energy flows of the remaining two concepts are explored more thoroughly in Chapter 3. Heat flow is studied first, since it is indicated by literature as the most significant loss mechanism for small scale engines. The characteristics of the complex real life situation are modeled in a multi physics simulation and linearized around the scale of interest for the FWM. The catalytic reaction is the primary energy inflow for all concepts, but based on measurements done by A.J.H. Meskers [60] it was observed that for Concept 1 the catalytic reaction characteristics might be critical. This is because the reaction time of small drops was found to be in the same order as the opted cycle time for the FWM in its current size. To study the characteristics of the catalytic reaction in more detail, a model is constructed based on the energy balance and fitted to the experimental data. The last energy flow identified is the exhaust fluid flow in concept 1. Due to time restrictions a full detailed analysis of this subject needs to be done in future work, but a basic analysis using the theory of gas dynamics is given. By approximating the shape of the exhaust as a round nozzle and assuming that friction has not much influence, the mass flow is given as function of a pressure difference. The findings of Chapter 3 are used in Chapter 4 to construct two models for the two remaining concepts. A numerical method for finding the steady state of these models is discussed and tested by comparison with analytical results. The performance of the models is obtained using these steady states. For Concept 3 it is observed that the performance is too low for the FWM in its current size, but has potential if the geometric scale of the system is reduced. This is due to favorable power density scaling of Concept 3. Of course this is only true if the operation principle of the real engine closely resembles the model. In the model a fixed temperature difference between the two heat sinks is assumed. This might be problematic in a real situation because it will be more difficult to accomplish a temperature difference when the geometric scale is reduced. For Concept 1, the engine model is a combination of the energy balance of the species inside the cylinder and the exhaust flow model described in Chapter 3. To reduce the sensitivity of the conclusions to the uncertainties of these models, two sets of parameters are used. One is considered optimistic and the other pessimistic. It was found that the performance for both sets is in the right region for the FWM engine requirements. This suggests that among the considered concepts, Concept 1 is the best way forward for the engine project of the FWM.