Powder based three-dimensional printing (3DP) allows great versatility in material and geometry. These characteristics make 3DP an interesting method for the production of tissue engineering scaffolds. However, 3DP has major limitations, such as limited resolution and accuracy, hence preventing the widespread application of this metho engineering. In order to reduce these limitations deeper understanding of the complex interactions between powder, binder and roller during 3DP is needed. In the past a lot of effort has been invested to optimize the powder properties for 3DP for a certain layer thickness. Using a powder optimized for an 88 mu m layer thickness, this study systematically quantifies the surface roughness and geometrical accuracy in printed specimens and assesses their variation upon changes of different critical parameters such as the moisture application time (0, 5, 10 and 20 s), layer thickness (44 and 88 mu m) and the number of specimens printed per batch (6 and 12). A best surface roughness value of 25 mu m was measured with a moisture application time (using a custom made moisture application device mounted on a linear stage carrying the print head) of 5 s and a layer thickness of 44 mu m. Geometrical accuracy was generally higher for the 88 mu m thick layer, due to a less critical powder bed stability. Moisture application enabled 3DP of a 44 mu m thick layer and improved the accuracy even for a powder initially optimized for 88 mu m. Moreover, recycling of the humidified powder was not only possible but, in terms of reactivity, even beneficial. In conclusion, moisture-based 3DP is a promising approach for high resolution 3DP of scaffolds. (C) 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.