LOSS COMPENSATION METHODS FOR STONE
JOHN GRISWOLD, & SARI URICHECK
6 INORGANIC BINDERS
There are three general classifications of inorganic systems: plasters, naturally occurring cements, and modern cementitious mortars. Despite the advent of modern polymeric materials, which offer a greater range of visual effects, inorganics are still chosen for stone fills mainly because of their strength, stability, durability, and availability.
In these fills, both binder and filler are composed of carbonate and silicate networks. The curing of these mixes is achieved by reaction of these constituents with water and oxygen. In these mixtures, many types of aggregate react chemically with the binder, rather than passively filling space within the binder matrix.
Disadvantages include excessive hardness; introduction of soluble salts; poor permeability by salts and moisture; shrinkage; introduction of large amounts of moisture to the substrate during use; and a generally cool, opaque appearance.
There are two main types of plaster: gypsum (calcium sulfate) and lime (calcium carbonate). Both have been used since antiquity for restoration of sculptural and architectural work. They are often the only component of a repair mixture, but they may be bulked with fill materials such as crushed stone and microballoons to enhance textural qualities. The color of a plaster repair may be integral (included in the curing mix) or added later as a painted surface finish.
Gypsum plaster, e.g., plaster of paris, has traditionally been used for restoration in ceramics and stone, and much has been published on its properties. Its basic chemistry is calcium sulfate hemihydrate, which takes on additional water on mixing and loses it as the gypsum network recrystallizes on drying. Industrial grades are commonly superseded by dental plasters; these plasters are in a refined state and are quick-setting due to added fumed silica, which allows application of the fresh mix on vertical surfaces. Patent plasters such as Keene's Cement, invented in England in the late 19th century, are modified gypsum plasters. The manufacture of such plasters involves repeated heating above 170°C to form the anhydrite of calcium sulfate. A seed catalyst such as alum is added during heating to enable the plaster to set (Ashurst 1979). Because it forms a tightly intertwined network of needlelike crystals on curing, it becomes harder and is less soluble than plaster of paris. It found widespread use in the manufacture of scagliola (imitation marble) and for architectural casts. Unlike plaster of paris, which cures quickly, Keene's Cement can be kneaded into a doughlike loaf, which remains pliable for hours if kept moist. Introducing pigment on the surface of the loaf, folding it, and slicing thin layers with silk thread or dental floss creates a three-dimensional facsimile of veined marble.
Many commercial plasters like Polyfilla (in the United Kingdom) and its counterpart Permafill (in North America) are also modified gypsum plasters. In their mixes are cellulose-based additives, which strengthen the mix and make it pliable for sculptural modeling, and whiting (calcium carbonate), which retards setting and increases the hardness of the gypsum plaster. A number of other additives have been used to modify the properties of gypsum plaster. To slow the setting time, glue, starch, and vinegar have been included in traditional recipes. Carbohydrates such as sugar or beer also slowed the setting rate and produced a harder plaster (Thornton 1992). Glues and polyvinyl acetate emulsion are also used to increase strength and reduce porosity of plaster fills. Common practice now includes wholesale consolidation of plaster fills with Paraloid B-72 (Koob 1987).
It has been argued that gypsum plaster introduces soluble salts to a substrate (Ashurst and Dimes 1990), but others feel that the use of refined chemical-grade plaster eliminates this concern (Soultanian 1996). Acrylic resin barrier coats on break edges or the attachment of cured plaster fills with an adhesive are used to reduce salt migration. Despite these concerns, as well as slight solubility after curing, plaster repairs are often used on stone sculpture in indoor contexts. They are one of the simplest and quickest types to fashion and finish. Coloring of plaster fills may be accomplished by painting after cure, by introducing pigments in the dry plaster powder, or by adding pigments, paints, or dyes to the wet mixed plaster. Reasonably acceptable faux finishes are achievable by widely practiced decorative painting techniques. However, the durability of such painted surface finishes on plaster, outdoors or indoors, is often unsatisfactory in the long term.
Lime plasters share a similar binding chemistry to gypsum plasters in that their curing is based on the mixture and subsequent loss of water, which returns the original compound before water was introduced. Lime technology is based on the burning of limestone (primarily calcium carbonate) to form quicklime (calcium oxide), which is then “slaked” by adding water to form the hydrated form, calcium hydroxide, Ca(OH)2. The mix sets on drying, and as carbon dioxide is absorbed the lime slowly hardens by reverting to calcium carbonate. For a concise review of chemical terminology and mechanisms pertaining to its manufacture, application, and curing, see Ashurt and Dimes (1990).
Lime putty, the product of mixing water and quicklime together, is often used as a binder. However, it is not a strong binder by itself and requires a filler such as sand or stone powder (Szczerba and Jedrzejewska 1988). It shrinks on setting and remains slightly soluble in water over time. Being weaker than cements, these lime plasters are compatible with ancient mortars and traditional stonework, whose strengths have diminished over time. While lime mixes may lose structural integrity due to the water solubility of calcium carbonate, this quality is preferred over the use of a repair that is too strong and poses the risk of damage to the substrate (Sass 1996). Custom-made lime mortar mixes are often chosen over commercial mortar mixes because soluble-salt content is easier to control, since the choice of raw materials may be monitored by the conservator.
Acid dissolution, difficult working properties, and the need for a dry-set sometimes preclude the use of lime plasters as fill materials. Monuments with moisture problems due to rising damp are especially poor sites for their use (Peroni et al. 1981). The dissolution products of lime plaster, including alkaline-soluble salts, can damage adjacent areas. However, because of their efficacy in other contexts, specialized techniques are constantly being developed for working with lime mortars, such as hammering components to induce microporosity and adding organic components such as straw (Ma 1995).
6.2 NATURAL CEMENT
Lime plaster is the main component in cements. A cement is a mortar that sets in the presence of water and is classified as a hydraulic mortar. Hydraulic mortars cure by the reaction of hydrated lime with silica and alumina components, which may be deliberately added or simply present in naturally occurring clays that contaminate the lime source. Limes that naturally contain a high amount of clay are called hydraulic limes and create a high-strength natural cement (Wisser et al. 1988). Added alumino-silicate sources may be termed pozzolanic additives. These are named after pozzolana, the high-silica volcanic ash from the Pozzuoli mountains, which so enriched the lime resources of Italy. Examples of pozzolanic additives include brick dust or fragments, charcoal ash, volcanic stone and ash, river sand, glass, and certain types of crushed stone. All impart greater strength and insolubility to the cured mortar.
Since ancient times, repair of buildings has been accomplished by the use of natural cements otherwise known as Roman mortars. Vitruvius describes them in De Architectura (VII.II.2) (Morgan 1960), and since the first century A.D. they have not changed drastically. The same hydraulic lime and pozzolana-based binder is utilized (though more refined than ancient compositions) with the addition of aggregates of crushed stone or glass and pigments.
Because cementitious fills shrink upon curing, a thickly applied layer will dry with extensive cracking. To compensate for this cracking in large fills, a two-layer system is used, with a coarse ground-primer layer (coccioposto), and a fine surface layer (stucco) above (Demitry 1988). A typical composition, as used for restoration of the Arch of Septimus Severus, is “for the coccioposto: 1 part slaked lime, 1.5 parts pozzolana, and 1 part brick fragments; and for the stucco: 1 part lime to 2 parts of fill which is: 4 parts marble dust, 3 parts river sand, and 1 part sifted pozzolana” (Nardi 1986, 5). In England, a somewhat weaker fill that utilizes the same chemistry is a common recipe incorporating a semihydraulic lime called Totternhoe lime, which is a “feebly hydraulic grey lime” (Wingate 1988, 9).
Much research has been devoted to determining and repeating the compositions of ancient mortars (Penkala and Zasun 1988; Ma 1995) with pozzolanic additives (Wisser et al. 1988; Penelis et al. 1989) in building and ancient restoration campaigns. Their success and durability over the centuries gives natural aging data by which to judge them. Studies of their properties and how to manipulate them through compositional changes have been pursued (Holmstrom 1981; Peroni et al. 1981; Szczerba and Jedrzejewska 1988; Ma 1995), though the applied results are not generally found in the literature. However, the ancient and natural chemistries of pozzolanic additives have been adopted through the introduction of modern hydraulic additives like siliceous earth, condensed silica fume, fly ash, and phonolite dust (Wisser et al. 1988).
6.3 MODERN CEMENTS
In the 19th century, Portland cement was designed to mimic naturally hydraulic cements. By firing clay and lime at high temperatures, an “artificial” hydraulic lime is synthesized. Typical mortar mixes are made of white and/or gray cements, hydrated lime, sand, and stone flour with alkali-stable pigments. Variations of the ratios (Weiss 1989) and choice of additional aggregate can modulate the texture and color of the fills. Addition of crushed stone or soil for color is considered to be better than using pigments, as the migration of unreacted pigments to the perimeters of patches is a common problem. For conservation purposes, Portland cement's strength, low porosity, high alkalinity, and soluble salt content often limit its use because they contribute to accelerated deterioration of the adjacent stone (Cassar 1988).
Changes in the cement/water ratio can aid in controlling the porosity and to some extent the strength of a mortar by dictating the free water in the mix. As the mix cures and this water evaporates, its volume becomes free space; hence, the greater the free water, the greater the porosity. However, this effect may be counteracted by the formation of a “float” of finer particles at the surface of the fill because of excess water, resulting in reduced porosity and other inhomogeneous properties.
Resin and natural additives have traditionally been incorporated in cements to enhance the adhesion to the substrate along the “bond line.” Organic polymers in the form of aqueous dispersions or emulsions are added to cement in commercial formulations to improve adhesion, modify water absorption properties, shorten or lengthen setting time, increase strength, and reduce shrinkage on drying. Other organic additives to concrete, such as fibrous elements (e.g., jute), microfoaming agents, and surfactants, induce porosity but do not ultimately contribute to the strength or appearance of the fill on curing. The addition of acrylic latex or PVA emulsion additives can increase strength and water repellency in grouting mortars and has frequently been adopted in patching formulations. Ancient recipes cite the inclusion of “blood, egg, sugar, cheese and dung” (Sickels 1981) as additives in mortars. The incorporation of alkoxysilane stone consolidants in the patching mix has been reported as well (Hempel and Moncrieff 1977; Andersson 1986). The inclusion of microfoaming agents, which encourage air-entrainment, is shown to modify porosity in natural cement mixes, but this approach does not necessarily create increased air or water permeability because an interconnected network of pores is not guaranteed. The addition of natural or synthetic microfibers may help provide such a network (Ma 1995).
The flexibility of these recipes is often cited as an advantage by those who utilize them. For some conservators, however, the versatility of such mixes is viewed as a flaw in the method. Accurately replicating a recipe consisting of many components is difficult, and for a large-scale or long-term project, slight differences between batches can produce a mottled, inconsistent repair. This challenge is one that commercial products have addressed.
6.4 COMMERCIAL MORTARS
Commercial products have been developed to make consistent results between batches easier to achieve. Some possess superior working properties compared to “homemade” recipes. The major disadvantage of choosing a commercial product is uncertainty about the ingredients due to the protections afforded proprietary companies, which allow the details of product compositions to remain secret. Despite this drawback, commercial mixes, with appropriate analysis, should not be ignored in cases where the consistency of results in a cementitious repair is vital.
A number of commercial products are available, though the selection of conservation-quality materials is more limited. Products are often favored based on regional biases of individual manufacturing centers and advertising focus. For example, Ledan and Mapei products are primarily manufactured and used in Italy. Keim is a German company whose sales efforts focus on Germany, Scandinavia, and Great Britain, but these products are also used in North America to a lesser extent. Jahn Mortars is a Dutch company that has an international clientele with a growing base in America through its sole distributor, Cathedral Stoneworks. Edison Coatings is an American product line whose market includes Mexico, Canada, and England. What distinguishes these products, in general, from other nonrestoration quality materials is lower alkalinity, lower soluble salt content, and lower strength. Some products are also custom matched by the manufacturer to visual and physical properties of specific stones.
This discussion is restricted to two commercial products available in the United States, Jahn mortars and Edison patching systems. Both of these product lines are in common use by conservators, but few references to their use have been made in the conservation literature (e.g., Wheeler and Newman 1994). In essence the basic cementitious lime-silicate chemistry is the bonding mechanism for both product lines, although Jahn Mortars include natural pozzolanas. Additives like microfoaming agents, which create porosity, may also be included in each, but this is not confirmed. As for all cementitious mixes, the curing schedule is a crucial part of a system's inherent properties, and Jahn and Edison provide specific mixing and curing instructions and suggestions. The properties for Jahn are stated in the product literature, and preparation and application parameters are aggressively encouraged through written instructions and mandatory training courses.
Each mortar mix is composed of sand, lime, and mineral additives, with the addition of crushed stone in the Edison line. A look at the raw components of each dry mix demonstrates this marked difference in the textures. Jahn Mortars appear as a fine grained powder, while Edison's Customs System 45 shows a texture with a greater variety of aggregate size. Discrete stone flakes can be easily discerned in Edison, while not in Jahn. This variance in raw material translates directly into different cured textures as well. Admixtures of other fillers can be incorporated, but the guarantee of certain physical properties will be lost. In addition, there are limits to the amount of bulking of a mix that can be tolerated in the network strength, and these limits need to be determined.
Both product lines offer the service of custom matching to a provided substrate sample. Jahn products are completely inorganic and, as such, circumvent issues of light stability of resin components. The products are dry powders to which is added a relatively small amount of water. Edison's products include an acrylic emulsion added to a dry powder and water mix. Although the acrylic is not part of the curing mechanism, it augments the strength of the bond between the patch and the substrate. Therefore, a thinner, more feathered application is reportedly possible than with the Jahn products. The potential disadvantages of the acrylic are a fill strength that may exceed that of the stone, decreased water permeability, and risk of discoloration and degradation with exposure.
For Jahn M70 and M120 products and the Edison Custom System 45, standard batches for limestone, marble, sandstone, granite, and other materials are modified to match the specific stone sample submitted. Matching is offered in color, texture, and physical properties, such as compressive and tensile strength, permeability, modulus of elasticity, and coefficient of thermal expansion (see product literature). Multiple mixes of hues may be ordered to accommodate the natural color variations in a given stone. Individual modification of the batches by the addition of pigments is feasible but not recommended by the manufacturers for most uses. Consultation with technical support staff has proven fruitful in addressing specific conservation treatment problems for objects.
According to the practical observations of several conservators, the strengths of the Jahn line include carvability and ability to build up large complex shapes in a single application. Reported shortcomings include their fine texture, which does not accurately match all stone, and their opacity, which means they cannot closely match a translucent stone. Since there is no added acrylic polymer, a minimum depth and keyed edges are needed for a Jahn fill to adhere effectively; feathering out the edges of the fill is not advisable. Edison products reportedly can have a slight sheen, which can be favorable or unfavorable depending on what type of surface is to be repaired. Diluting the additive or redressing the patch surface can reduce this sheen and is usually considered a minor concern (Champe 1996).
In spite of strict recommendations for the preparation and application of these commercial products, conservators have found ways to modify their properties to suit particular requirements of deteriorated stone. For example, different ratios of mix to water will yield varied properties (Williams 1996). For degraded local areas of a particular stone, the commercial patching mixes made specifically for that stone can still be too strong. Conservators use several ways to reduce the strength of the mortar in the field. Adding crushed or powdered stone to the batch, adding excess water in the mix, adding water to reconstitute a partially cured mixture in the bucket, or skipping the recommended regimen of misting the patch with water after application can all result in a weaker fill. The results, however, are not easy to predict or control.
The inorganic mortars described above are almost exclusively used on stone treatments in outdoor environments. Rarely are they used as a repair method for museum objects, where such strength and weather resistance are generally not required. Ashurst and Dimes (1990) note that because hydraulic limes are too high strength and impermeable, and transfer soluble sodium salts, cements are excluded for stone sculpture. In architecture, appearance is often sacrificed for durability, while in sculpture, appearance is often paramount. Inorganic mixes are sometimes difficult to incorporate aesthetically with a substrate. Emulating smooth and translucent stone is always a problem. Even if the initial match is successful, it may become differentially soiled in comparison to the original stone, particularly around the edges. The wet curing conditions required for these mortars contribute to their exclusion from museum conservation because the wetting of the treated object is often not advisable. Other disadvantages are the soluble-salt content, shrinkage, incompatibility due to differential thermal expansion, and the need for aggressively interventive surface preparation. Obviously, the relative risk that these factors pose, weighed against the benefits of their usage, must be assessed before deciding to use an inorganic repair mortar.