Many artifacts and works of art are too fragile to be accurately replicated for scholarly study by using traditional methods of plaster or silicone molding techniques. In addition to the potential for loss of surface and fracturing, the use of many modern silicone mold materials may lead to irreversible staining of the original surface.

An evaluation of newly developed rapid prototyping technologies and their potential application to this problem was conducted by the J. Paul Getty Museum Department of Antiquities Conservation in conjunction with Laser Design Inc. of Minnesota and Hughes Aircraft Corp. of California.

Rapid prototyping technology allows accurate copies to be made without surface contact with the original object. This paper discusses the outcome of an initial investigation into one of these processes, stereolithography. A plaster model was submitted for laser scanning using a point laser probe, programmed for a 3.175 mm (0.125) distance between scan lines. The input was filtered and stored in an STL (stereolithography) format which provides the X,Y and Z coordinates of the A,B,C and normal vectors of a predetermined number of surface triangles. This data was then used to produce a three-dimensional copy.

Stereolithography produces hollow or solid threedimensional forms by feeding data that make up cross sectional slices of the object scanned to a focused laser aimed along the Z axis of the potential copy. The laser polymerizes the outer contour of the 'slice' in a thin layer of photopolymer deposited on a stage located below the laser. Following each slice production the stage lowers approximately 19 mm (0.75) below the surface of the liquid polymer reservoir and upon resurfacing positions approx 0.05mm (0.002) lower than its previous position. A “squeegee” gently crosses over the surface resulting in a thin film of polymer available for the production of the next cross section. The laser is then sent the next cross sectional data and the sequence of section building continues.

The specific challenge presented by this project was to assure that the surface detail of the copy formed presented acceptable fidelity to the original object (model), avoiding any need for follow up bench work or correction. The advantages, drawbacks and newly developed alternatives to this promising technology are discussed.