EDDY CURRENT MEASUREMENT OF DENSITY DURING HOT ISOSTATIC PRESSING H. N. G. Wadley, A. H. Kahn, Y. Gefen, and M. Mester· National Bureau of Standards Gaithersburg, MD 20899 INTRODUCTION Hot isostatic pressing (HIPing) is an increasingly used process for consolidating and densifying metal and ceramic powders to near net shape. Powder is encapsulated in a thin walled cannister under vacuum and placed in a pressure vessel where it is subjected to a temperature/pressure cycle, Fig. 1. The cycle used is normally empirically determined and aims to achieve 100 percent theoretical density in the sample. Because of the expense of HIPing, the empirically derived HIP cycles are often not optimized. Changes in part geometry, powder composition, and size distribution, etc., alter the pressure/temperature needed to attain full density. Thus, some cycles may result in less than full density while others may subject a component to more extensive heating than that required for complete densification. The former leads to very poor quality components while the latter adversely affects process productivity and can result in unacceptably coarse microstructure for some applications. One possible approach to the problem is being explored by M. F. Ashby and co-workers at the University of Cambridge [1,2,3]. This work is attempting to model the consolidation and densification of powder and to use the models to predict densification as functions of pressure, temperature, and time. These model predictions can be conveniently represented in the form of HIP maps (described further below) and could provide a theoretical basis for selecting optimum HIP cycles. The models have had little experimental validation to date because of the time consuming and expensive nature of interrupted testing and the lack of an in situ method for measuring density. Even if the models do accurately represent densification, their value is limited because predicted density depends upon intrinsic properties of the powder (e.g., yield strength, power law creep hardening exponent, etc.). These data are sparse and inaccurate for many potential systems of interest. The models could be used to infer the necessary data from measured densities if again an in situ density measurement technique were available. *Research Associate, The Aluminum Association 1589