In this paper we thermodynamically assess the performance of an ammonia-water Rankine cycle that uses no boiler, but rather the saturated liquid is flashed by a positive displacement expander (e.g., reciprocating, centrifugal, rotating vane, screw or scroll type expander) for power generation. This cycle has no pinch point and thus the exergy of the heat Source can be better used by matching the temperature profiles of the hot and the working fluids in the benefit of performance improvement. The second feature comes from the use of the ammonia-water mixture that offers further opportunity to better match the temperature profiles at the sink level. The influence of the expander efficiency, ammonia concentration and the coolant flow rate is investigated and reported for a case Study. The optimized cycle is then compared to four organic Rankine cycles and a Kalina-type cycle and shows the best performance. It is also shown that, in order to determine the best cycle configuration and parameters, energy efficiency must be used only in conjunction with the amount of the heat recovered from the source. The efficiency of the cycle running with ammonia-water is 0.30 in contrast to steam-only case showing 0.23 exergy efficiency, which means an increment of 7.0% is obtained for the same operating conditions. If cogeneration is used the cycle effectiveness may even be over 70%. The cycle can be applied for low power/low temperature heat recovery from geothermal sources, ocean thermal energy conversion, solar energy or process waste heat, etc. (C) 2008 Elsevier B.V. All rights reserved.