A cut above the rest

Gem cutting has always been the preserve of craftsmen, but two research teams hope to automate the process to minimise lost material and extract the maximum value.

In 1673, the jeweller to the French court, Sieur Pitau, was given the task of cutting a triangular lump of blue diamond weighing over 112 carats. When he had finished, the faceted jewel weighed just 67 carats — almost half had been lost.

Over the years, the stone has been recut, shrinking it even further. Its remaining fragment — the Hope Diamond, in the Smithsonian Museum, Washington — weighs a little over 45 carats. The stuff of legends, but a nightmare for a gem merchant.

It’s inevitable that material is lost when transforming a rough precious stone into a finished gem. Cutting has always been the preserve of craftsmen, but engineers are seeking to automate the process. Machinery and computers, it seems, might be the way to minimise the amount of lost material and extract the maximum value from the stone.

At Germany’s Fraunhofer Institute, Karl-Heinz Küfer is leading a project to produce a fully-automated gem-cutting plant. He is developing software which will measure the rough stones accurately; decide what shape stone would fit best inside the rough shape; and control grinding and polishing machinery to produce the final stone.

‘Common industrial processes are designed to make standardised products. Here, though, we are talking about high-precision, individualised production, because no two uncut stones are the same shape,’ said Küfer.

The system starts with accurate measurement of the stone using moiré projection, where thin, straight strips of light are projected on to the stone, and their curvature as they pass over the surface analysed to give a computer model of the shape. The software fits a variety of designs for the cut gem inside the rough shape, and compares the current market prices for the various cuts to decide which shape would extract the maximum value from the stone. This data is then used to control cutting and polishing robots which, according to Küfer, are accurate to within one micrometre and a thousandth of a degree.

The software and machinery are currently the subject of a patent application so Küfer cannot discuss them in detail, but a prototype plant in Idar-Oberstein, an established gem cutting centre, is scheduled to begin operating next year. Küfer’s technique is suitable for rubies, emeralds and sapphires, but not for diamonds, which depend on the precise orientation of the facts with the natural crystallographic planes of the crystal for their sparkle.

But at Cambridge University, Tony Holden and Matee Serearuno are working on a system which could form the basis for fully-automated diamond cutting.

The value of a diamond depends on its carat, colour, cut and clarity. Three of these are easy to determine, but the last — the number, size and position of imperfections in the stone — depends on the decision of a skilled assayer.

Holden and Serearuno have devised software called iGem which automatically grades a stone, using fuzzy logic algorithms based on the rules used by expert assessors. It then suggests which cuts would maximise the stone’s value, or even improve its grade, for example by removing surface imperfections.

The researchers are working with a Johannesburg firm, Calibrated Diamonds, to incorporate iGem into an automated diamond cutting system.

‘We can use lasers for the surface scanning, and X-ray tomography for the internal imperfections,’ said Holden. Calibrated Diamonds already uses lasers to perform the rough shaping, or bruting, of the stones. ‘Integration of iGem with CAD “shapefitting” software would allow the cutting pattern for a rough stone to be designed, taking account not only of its parameters, but also what shape of finished stone a customer wanted,’ said Holden.

‘Once the design was made, a computer-controlled robot or laser cutting tool could perform the cutting task automatically.’