Fall of man reversed

Materials World magazine
1 Jan 2015

Simon Frost looks at research into adhesives and materials in the restoration of a severely damaged Renaissance sculpture. 

A 15th Century marble sculpture of Adam that shattered into 28 pieces and hundreds of tiny fragments in a fall has gone back on display at New York’s Metropolitan Museum of Art (Met), after years of unprecedented collaborative research and repair by conservation scientists, materials scientists and engineers.

The statue was carved from a block of Carrara marble by Italian sculptor Tullio Lombardo in the 1490s. Its plywood display pedestal gave way in October 2002 under 190cm-tall Adam’s 350kg weight, leading conservators on a 12-year mission that posed a unique set of challenges.

Picking up the pieces

The first step was to laser scan the major fragments to create computational 3D models and virtually reassemble them. Critically, CAD modelling allowed a team of engineers to conduct an advanced finite element analysis – for which no studies existed in art conservation applications. Computer Aided Engineering Associates (CAE) of Middlebury in Connecticut, USA, examined how compressive, shear and tensile stresses would be transmitted across the fracture surfaces.

CAE gathered results from three studies – a faceted stereolithographic (STL) model, a smooth non-uniform rational basis spline (NURBS) model, and an innovative hybrid study, where the rough shape of the broken surfaces in the STL model were isolated and imported into a smooth NURBS model for definitive accuracy. The maximum compressive stress was located at the base of the left calf, tensile stress at the back of the same fracture, and shear stress at the connection of the hip and torso.

With the stresses identified, the team set about evaluating the best adhesives and pinning materials for repair. They examined the thermoplastic adhesives Paraloid B-72 and B-48N because of their desirable chemical stability and reversibility. B-72 is commonly used in conservation as a barrier coating between fracture surfaces and thermosetting epoxy or polyester resins, to make the joint reversible. The Met team was the first to conduct research into its qualities as a standalone adhesive and in combination with B-48N. Collaborating with Columbia and Princeton Universities, the team examined the interfacial fracture toughness, bond-line thickness, adhesive and solvent retention, and creep behaviour of B-72, B-48N and varying combinations of both.

The tests showed, for the first time, that thermoplastic adhesives can be used independent of irreversible thermosetting glues in structural repairs. The adhesive with the best combination of reversibility, strength without creep, minimal bond line and ageing characteristics was found to be a 3:1 blend of B-72 and B-48N.

Common practice in conservation favours stainless steel or titanium pins coupled with adhesives, but this combination was much stronger than was required to sustain the loads involved. Therefore any failure would damage the surrounding marble, rather than separating at the join line as the pin fails.

In one set of tests, Carrara marble cylinders were cut at 45 degrees across their centre to mimic the shear joins of Adam’s ankles and left knee. The two halves were joined together through a central drill hole with steel, titanium, carbon fibre and fibreglass pins, then subjected to gradually increasing compressive force in a mechanical analyser until either the pin or marble failed.

Fibreglass pins were chosen after exhibiting sufficient strength to withstand the maximum static forces of the sculpture, while causing no damage to the marble when pushed to failure – unlike commonly used metals.