Experiences in anchoring systems in the restoration of stone artefacts by Guy Devreux + Stefano Spada

User menu

The Vatican Museums Studio for the Restoration of Marble and Casts is working on new solutions so as to keep on improving mechanical anchoring systems for detached parts of stone artefacts. In essence, the aim is to make these systems as reversible as possible, in some cases modifying the philology of the procedure. This paper describes three different methods to be used in three different situations. The first uses a pin combined with a hole which passes through one of the parts to be attached, the second uses a plain pin inserted without using adhesives ¬to be used only in situations not involving traction forces – and the third provides a solution for attaching suspended or semi-suspended elements. In this latter case, the pin is only partly involved in anchoring the piece: primarily it acts as a safety bolt which impedes the complete detachment of the reattached component to prevent it from falling off. All the systems described were designed mainly to facilitate adhesion methods, to improve their effectiveness and to be reversible.
In recent years the Vatican Museums Studio for the Restoration of Marble and Casts has dedicated itself to advancing and developing treatment methods, leading to new and substantial changes in its procedures. The object of this has been to make treatments more effective, but above all to make them more respectful of the artworks and to make them as far as possible fully reversible. The subject under discussion hereis central to a significant part of our work ¬ the reassembly of sculptures – and involves the systems of mechanical connection (pins) used between detached components.
Traditionally in the restoration of stone sculptures, when parts needed to be joined together, combined with the use of adhesives, there was often recourse to the use of one or more pins or rods – or some other anchoring elements such as brackets, clamps, holding frames, or pins or rivets made from molten lead – to attach parts projecting to a greater or lesser extent from the sculpture. In various different ways, wax-resin, plaster and lead were employed both as adhesives as well as fillings, and thus applied both on the fracture/join surfaces and around the mechanical joining element (pin or rod). Even today, when reattaching a part, particularly when there is an old hole available, the general practice is to apply a mechanical element and an adhesive. These form part of a “joining system”. Nowadays, as opposed to earlier practice, we try to preserve the distinction between the mechanical and the adhesive components of this system, maintaining them as independent in their role and function as possible.
The traditional adhesive which we most often come across in the re-restoration of ancient sculpture is the well-known “mistura” or “mestura” – a mix of beeswax and colophony resin (“pece greca”), sometimes with aggregates such as marble powder, brick dust, etc. This adhesive, whose properties warrant more detailed research, has excellent reversibility especially if compared to many modern synthetic resins such as epoxies and polyesters, as it dissolves in organic polar solvents such as ethanol and acetone even after some centuries. The biggest disadvantage of mestura is that over decades, if not centuries, it suffers both aesthetical and mechanical degradation. The adhesives available today are much more resilient and they are much stronger both internally and at their bond surface than the cohesive inter-granular strength of sound marble, even if applied in very thin layers. The excessive mechanical properties of an adhesive can be a drawback, not only from the reversibility point of view, but also from the safety of the joint itself. For, if a repair (for example, in the typical case of a statue with a broken arm) made with these kind of adhesives was subject to a serious physical shock, any break would occur in the intact marble, close to the old repaired fracture/join rather than in the adhesive, producing new damage to the object.
Ideally an adhesive for the joining of stone elements should possess mechanical characteristics slightly inferior to those of the treated stone in the area around the joint. Another problem is that adhesives like epoxies and polyesters still have poor reversibility and, typical of polymers, undergo gradual alteration of their appearance and mechanical properties over the medium to long term. While we are hopeful that an ever greater variety of improved adhesives will be developed, our approach at present is to use the minimum quantity of adhesive necessary, applying it, unlike in the past, only to the stone interface of the joint which has been previously coated with a reversible intervention layer using Paraloid B72. This intervention layer is also applied in advance to the surfaces of the seats of all pins, cramps, etc. to be filled later. The type of filler applied around the pin or other mechanical element depends upon the specific nature of each case; however, we seek as far as possible to avoid the direct application of a material with adhesive properties to the mechanical element itself, using instead a material with filling properties and sufficient compactness. Considering all these factors, that an adhesive should be weaker than the treated stone, that it alters over time, and that it should normally be used in minimal quantities (often necessarily so as the voids available are less than a millimetre in size), it seems even more important to clarify the function of any pin or mechanical anchor which may be used.Where heavy weights are involved, and differently orientated relative load vectors, and in the presence of already existing holes (and at times when it is necessary to drill a new hole), it is generally best to use a system of reversible pin + "weak" adhesive rather than an "ultra-strong" adhesive without a pin. The pin, in fact, should have the function of connection and load distribution in relation to the masses involved, as well as assuring safety should the adhesive fail, either naturally or intentionally, while planning should include as far as possible reversible treatments.
The development of composite materials and their gradual and cautious entry into the field of stone restoration, supported by planning and testing using techniques such as Finite Element Analysis applied to the 3D modelling of a sculpture and the fragment to be joined, allows us in some cases to reduce the dimensions of the mechanical elements to be used (especially, but not limited to, the diameter), while still providing the required mechanical properties, thus reducing the impact on the sculpture. This further encourages us to favour the solution "new pin + weak adhesive", rather than "strong adhesive without pin", albeit in the case of loads above a certain level bearing on a unit join area. On the other hand, below this limit, reversible adhesives such as acrylics, vinyls, methyl methacrylates, polyvinyl butyrals and similar are sufficient.
In accordance with this philosophy, in the Vatican Museums Studio for the Restoration of Marble and Casts we have recently been attempting to produce tailor made anchoring systems for each specific case. Factors to evaluate include the orientation of the linking pin, the shape and position of the piece or fragment (projecting on one side, locked in place, etc), whether there is a perfect join or otherwise, and whether the piece will be self-supporting and connecting or just self-supporting.
As part of a series of varied cases and their solutions that have been published by colleagues in our sector in recent years [1,2,3,4] and which seem to converge as examples of real "teamwork-in-progress", we would like to offer here our further constructive contribution. The following examples will clarify the concepts described above.
The first system, the simplest, is that of anchoring two elements using a pre-existing hole which opens onto the exterior of the work. In this case it is quite simple to create a thread, with the glue or with the insertion of a bushing into the hole passing through both the pieces to be joined; the threaded supporting pin is screwed into this and held in place at the top by a concealed blocking system that can be easily removed (Fig. 1).
The second system, far more frequently applicable, is used when there are stone blocks to be joined which only require mechanical reinforcement. Generally speaking, a pin is applied at right angles to the break we need to repair. This second “dry” system (without adhesive for the pin) can be used when the pin reinforcing the piece to be attached is at any angle between twelve o’clock (vertical) and three o’clock (horizontal). The pin is inserted into an anchoring system which is solidly attached to the parts which are to be joined. The adhesive is applied only on the contact surfaces of the join and not in the seat of the pin, making the treatment much more reversible. To ensure the success of this system, ideally there should be a smooth male-female connection for the pin. Thus before inserting the pin in the stone, the hole in the marble should be prepared beforehand by coating with synthetic resin to make it smooth (Fig. 1b) Even better, the seat for the pin is provided by gluing a corrosion resistant metal sleeve into each of the holes in the two stone pieces to be joined. In order to make this system more reversible, we have designed steel sleeves with threads on their exteriors and internal bottoms so that they can be more easily removed from the stone using a tailor made extractor (Fig. 2). All of these systems mean that anyone who has to dismantle our work just needs to break down the adhesive applied between the areas of contact to separate the join and extract the pins applied 'dry' (without adhesive). Additionally, in the procedure illustrated, in order to prepare two new holes for pins which needed to be perfectly parallel, a template was used which provided a guide not only for the position but the alignment of the holes. This template may be considered a doubled version on the horizontal plane of the "forassometro" - hole template - devised by Usai [2]. If required, the device can be used several times, starting from an existing hole, to create additional parallel holes.
The third case is when a part is attached with a simple join, that is when there are no other obstacles around the join (which is coaxial), there is a single mechanical linking element, there is no through-hole opening to the outside, and the pin and its hole are completely hidden inside the stone of the sculpture and may be accessed only when the attached piece is removed from the sculpture. In these cases, and where there is no other join at the other end of the attached piece, and above all when the part is inclined downwards, sloping below the horizontal line between 3 and 6 o’clock on our reference clock, we use a mechanism we have recently designed. It is an "anti-falling" system which ensures the piece is held even if the adhesive gives way (accidentally or intentionally), while allowing easy removal. This system, in terms of safety, is desirable above all for works at risk from vibration (works being transported for exhibitions, or in seismic risk zones) (Fig. 3). At the end of the pin there are two threads with opposite twists. Each one is fits into the threaded bushings previously correctly positioned using the centring rod and glued into place (Fig. 4). The pin can then be screwed simultaneously into both bushings and can be extracted by reverse rotation (Fig. 5). The ends of the centring rod and the pin have holes in the centre to enable the entry of two corresponding studs in the base centre of the threaded bushings, so that the two bushings can be perfectly aligned during installation, and thus also the pin during final assembly. The central part of the bolt has a thicker section equivalent to the internal diameter at the entry of the bushings, with the minimum tolerance necessary for ease of entry. This is to keep the pin as best aligned as possible once it is load-bearing and to increase the contact and load bearing surface between the pin and the mass acting upon it (stone-matrix-filler-bushings). The hole through this thicker section allows a pin-key to be inserted to screw and unscrew the pin. The pin should be screwed all the way to the end of the threaded section (Figs 5 and 6). Once the adhesive is added, just to the surface of the join (Fig. 7), the pin can slide freely for a further short distance until the two sides of the join fit together perfectly.
Reversing this process, to disassemble the system, the adhesive is broken down and the parts are partially separated. The pin can only move a short distance before being blocked by the contact between the male threads on the pin and the female threads on the bushings (Fig. 9). This is the greatest advantage of the system, as it guarantees secure hold on the pieces even if unsuitable adhesives are employed and/or there is stress or impact.
Another factor is, as in the second case described above, the external part of the bushings also have a thread, but shallower and wider than the internal one. One the one hand, these threads increase the roughness of the bushings’ outer surface, and therefore provide a greater surface area for the filling material within the hole to attach to; on the other, they are also intended to improve the removability of the bushings. The method used is as flows: the twist of the external thread on the bushing insert is opposite to that of the internal thread, so that if an extractor bolt (E and F in Fig. 3)is screwed into the empty insert until it reaches the end of the insert (Fig. 10),continuing to screw on it in the same direction will cause it to couple with the bushing and start to rotate the bushing in a direction which due to its external threads tends to draw it out of the hole (Fig. 11). It is, however, important that the bushing are lubricated but with appropriate materials: the type of lubricant used and its method of application must not fill the grooves on the exterior of the bushing, as this would dangerously reduce the bushing’s adhesion to the hole (previously deliberately improved by the threads). We used a fluorinated lubricant which is relatively nonreactive to resins. After lubrication, adding a very thin layer of tin or aluminium foil over the inserts will further aid removal. With the assistance of the Mechanical Workshop of the Governorate of the Vatican City State (the director, Claudio Cellante, and staff members Roberto Pietroletti and Sandro Sanità), the first prototypes were made in AISI 303 stainless steel.
It is crucially important to bear in mind that the mechanism described - pin/ bushings /filler/marble - works coherently with the stone surrounding it only if the bolt, with its widened middle section, is in contact with the insert (this thickened section may be made longer than in the first prototypes), offering resistance to the traction force exercised by the mass of the stone connected to it via the vector of the vertical component.
In other words, when released, at first the pin acts like a "dry pin" when it is moving freely. Thus, especially for loads which are heavy in relation to the area of the join, it would be better to have pins which are aligned from the horizontal to 30-35° below horizontal (between the hours of 3 and 4.30 on the clock). At the same time, it is important to ensure that the marble or stone around the joint is in good condition and has all the necessary mechanical properties to resist the traction forces of the adhesive applied there. In the absence of these conditions, we should rather consider the use of mechanical joins which are better anchored, less subject to slippage and in greater contact with the surrounding masses. We are working on a variant of the system described above which involves the use of a (partially elastic) block on the pin when in the closed position and, as a consequence, active collaboration of the pin in "holding" the joint along with the adhesive. This would employ springs with strong and weak compression, and a different design of certain details which we hope we can discuss on a future occasion. In this case we will be able to use the bolt with greater confidence in positions between 35 and 90 degrees from the horizontal (4.30 and 6 o’clock on the clock face).
Further developments of this system will involve testing its feasibility and efficiency when using other materials such as titanium and composites to improve the mechanical characteristics and further reduce the size of the components (especially the diameter, but also the length) while maintaining the parameters achieved by the original steel mechanism. Naturally in this case the threaded surfaces of the inserts would be maintained as they improve their adherence, indeed the adherence may be improved using fibre-reinforced plastics. Other possible improvements, particularly where heavy stone parts need to be joined (therefore requiring the use of longer pins) might be made to the extractor, which would have to be able to release longer and wider inserts: one method might be by attaching the extractors more solidly to the inserts, thus allowing the application of more torsion to remove them. This could be achieved, for instance, by making the extractor with a cylinder with a toothed rim which would insert into notches created along the external edge of the inserts, which would then be blocked with a nut that could be screwed onto the upper part of the extractor itself. The moment of release (intentional or accidental), would involve a short movement of the pin inside the sleeve followed by the impact of the pin’s and the bushings’ threads (Fig. 9) and could have cause damage to the internal thread, as well as on the stone due to the induced stresses in the zone around the joint. For this reason we are studying the possibility of inserting a shock absorbing component in the circular crown inside the inserts at the point where contact takes place. A first proposal involving the use of O-rings in Sorbothane (Sorbothane Inc.) was rejected due to difficulties of compatibility between rubbery thermoplastic materials inserted into a completely mechanical system intended for long term duration. We are studying other options, such as Ondufil (Borrelly) compression washers/springs or similar materials, of a suitable diameter so they would be fitted at the point inside the inserts where the two threads come into contact, and inserted in such a way as to enable the easy tightening and unscrewing of the pin.
In conclusion, the procedures we have so far succeeded in applying and testing are already partly applicable to and useful for the modifying our future treatments, both from a technical point of view and in terms of the philological aspect of the procedure itself.
Acknowledgements: Images ©Kavaklik, Rome
1. Bertorello Carla, “Un sistema di vincoli mobili per l’assemblaggio di sculture lapidee frammentarie”, in Materiali e strutture, L’Erma di Bretschneider, Rome, Year 2, No. 2, 1992, pp.67-72.
2. Usai Carlo, “Strumenti per eseguire fori di precisione su sculture lapidee”, in Materiali e strutture, L’Erma di Bretschneider, Rome, Year 2, No. 2, 1992, pp. 73-79.
3. Serino Carlo, Iaccarino Idelson Antonio, “Perni per l’assemblaggio reversibile di manufatti frammentari” in IX° Convegno Internazionale Scienza e Beni Culturali, Bressanone, 1-4, July 2003, Libreria Progetto Editore, Padova, 2003 , (with furtherbibliography).
4. Mamone Luigi, “Rimontaggio e restauro di statue lapidee frammentarie. Alcune esperienze ‘ sul campo’”, in Materiali e strutture, L’Erma di Bretschneider, Rome, Nos.11-12, 2008, pp.92-121
Guy Devreux is the Director of the Vatican Museums Studio for the Restoration of Marble and Casts, Vatican Museums, 00120 Vatican City, tel. 0669885295, e-mail lrm.musei@scv.va
Stefano Spada is Restorer at the Vatican Museums Studio for the Restoration of Marble and Casts, Vatican Museums, 00120 Vatican City, tel. 0669885295, e-mail lrm.musei@scv.va