Jul 182010
 
Decorative Plastering in Devon Image

During the 1960s and into the 1970s heritage bodies and organisations were becoming increasingly concerned at the accelerated decay of the overall fabric and detail of many of our churches, cathedrals and historic buildings. The three main culprits were environmental pollution, the extensive use of cement mortars since the 1860s and the presence of ferrous fixings throughout these complex structures.

By the 19th century the increase in pollution was rapid, due to an explosion of industry and population in cities, towns and countryside, and the consequent acceleration in the burning of fossil fuels. Limestone, as well as being a porous building material, is rich in calcium carbonate. This makes it highly susceptible to atmospheric sulphurous acids from the burning of fossil fuels. The calcium carbonate binder in the stone undergoes a chemical change to become calcium sulphate, which is one of the soluble salts most closely associated with the decay of the limestone through repeated cycles of salt crystallisation. By the middle of the19th century some of the carved surfaces of our historic buildings had accumulated a thick carbon deposit behind which the soluble salts were wreaking havoc. The result was the exfoliation and blistering of surfaces behind which the stone had been reduced to powder.

Hot on the heels of environmental pollution came the invention of Portland cement in the first half of the 19th century. This new and exciting material was highly regarded for its hydraulic properties, fast setting times and great strength. It is easy to see why the custodians of our churches and cathedrals must have thought that Portland cement was the ideal repair material which would cope with the increase in stone decay due to environmental pollution, particularly as it had the added benefit of colour matching the blackened surfaces. However, Portland cement is now known to be a potential source of sulphates, thereby contributing to the decay process. Of even greater significance is the fact that Portland cement has a dense and impervious nature which encourages concentrations of moisture and soluble salts at the interface of the cement repair and the limestone. In most cases where a cement repair is removed, the limestone beneath will be found to have been reduced to powder by the crystallisation of salts within the interface.

Jun 052010
 

Decorative Plastering in Devon Image

Lime’s early origins

Perhaps the earliest known examples of

decorative plasterwork are from the Old Kingdom in Egypt. Painted plaster

masks adorned the linen wrapped head of a mummy, and stone walls would have had their

irregular surface smoothed with plaster before being carved or shaped and painted. This

plasterwork was formed with fast setting gypsum plaster.

Roman stucco work, though mainly painted, shows widespread use of lime plaster, for

example; as a wall covering for landscape painting, as can be seen in Hadrian’s villa in

Tivoli in the 1st century AD; or as a theatrical backdrop of mythological figures and

theatrical figures in the upper class Hang houses in Ephesus, c 5th century AD.

Instructions by the Roman architect Vitruvius on the means of ensuring that stucco relief

decoration remains sound and firmly attached to the wall are as relevant today as they were

in the 1st century BC. His advice on the need for cane and metal support for relief work to

prevent distortion is, of course, common sense, as are the rules he describes for obtaining

a flat wall surface using three coats of plaster: a coarse base coat of rough sand and lime

reinforced with hair to prepare the wall surface, followed by a levelling coat of medium

graded sand, lime and hair to level the wall, and finally, a finer finish coat, much thinner

than the rest, of fine lime, sand and possibly goat hair.

Vitruvius’ advice on how to make lime plaster adhere to a damp wall has a particular

resonance today. To combat wet conditions, he recommended a pozzolanic additive of brick

shards and brick powders for the first of the three layers of lime plaster. The combination

of brick and lime, well mixed, provides a hydraulic set for the plaster (‘hydraulic’

literally means having the ability to set under water), enabling the mortar to set whilst

still wet, without carbonation.

In addition to brick dust, a multitude of other additives were used to accelerate the set of

lime, but perhaps the one ingredient that carries the most historical significance must be

marble flour. This aggregate was the key ingredient of the finest mid 18th century plaster

work, Stucco duro, which was largely confined to Italy and southern Europe. Marble

flour allegedly aids both the plasticity and the set of stucco. Although it was never widely

used as an additive by English plasterers, the style of the stuccodurists was much admired

and imitated. The twists and turns of a fine Rococo ceiling, with all its convoluted curves

and intertwining shapes, could not have been easily made without the setting properties of

marble dust or, as was later discovered by the English imitators of the stuccodurist, a

lacing of gypsum plaster.



Mar 152010
 

LIME

To make lime plaster, a limestone of almost pure calcium carbonate has to be chosen. This is fired in a limekiln at a temperature of about 1,000°C. The burnt stone taken out of the limekiln is quick lime (calcium oxide), a very caustic material that is difficult to keep, so it is almost immediately turned into lime putty (calcium hydroxide) by adding water, a process known as ‘slaking’ which generates a great deal of heat and steam.

Putty lime will harden slowly when exposed to air as the lime reacts with carbon dioxide to form calcium carbonate once again – a process known as ‘carbonation’. Fresh lime putty is therefore protected from hardening by being stored in waterproof containers in a damp state, permanently covered by a thin film of water.

Lime can be used by a mason to bed stones or modelled by a sculptor once the necessary aggregates have been added. (In plasters, aggregates such as sand are added in the proportions of up to around three-to-one for all but the finishing coat, principally to reduce shrinkage.) A modeller using lime plaster, or ‘stucco’ as it is often known, has time to change his mind some time after he has used it, for lime plaster will set over a five to ten day period. During this period it must be protected from drying out too quickly or it will crack. Once set, stucco will last for centuries.

 GYPSUM

Gypsum plaster behaves very unlike lime plaster. It is made simply by heating gypsum rock or alabaster – both of which are mineral forms of hydrated calcium sulphate – and grinding the result to a fine flour-like powder. At a relatively low temperature some of the water which makes up the crystalline mineral structure is driven off, forming calcium sulphate hemihydrate, which is then ground to a fine powder.

Gypsum plaster will set rapidly – within 15 minutes once it has been ‘knocked up’ with water – forming interlocking crystals of gypsum. This is not a material for modelling with, more a material for casting with, as it sets so quickly. So we have two completely different materials, for different purposes. A slow-setting lime plaster and a fast-setting gypsum plaster.

One of the earliest and most renowned sources of relatively pure gypsum rock was Montmartre, Paris, from which the material takes perhaps its most common name, plaster of Paris.

Plasterers, particularly since the late 18th century, have generally used gypsum plaster both to imitate earlier lime plasterwork and to create their own contemporary plasterwork of varying quality.



Mar 152010
 

The earliest structures to which plasters were applied took the form of panels of woven hazel or willow spars supported by timber. When first applied, some of the plaster would protrude through the spars, creating interlocking ‘nibs’ in the void behind. The nibs help to secure the plaster to the lattice, reinforcing the key or bond between plaster and wood. For centuries hair and other fibres have been added to lime and gypsum plasters to give greater strength to these nibs and stop them breaking off.

For the reinforcement of lime plasters and renders, hair should be strong, long and free from grease or other impurities. Ox hair is the preferred choice, but horse, goat, donkey, and a variety of other hair, including reindeer, are suitable. Human hair, being relatively fine and of poor strength, should not be used.

Traditional alternatives to hair include chopped straw, reed, manilla hemp, jute, sisal, and even sawdust. Modern synthetic fibres such as glass and polypropylene which have been designed for use with Portland cement mortars have also been used successfully in pure lime mortars, despite their smooth and almost shiny appearance when viewed under a microscope. Natural animal hairs, by comparison, have a much rougher texture, and are generally more appropriate for historic buildings.

While woven hazel or willow spars work well and are often found in surviving wattle and daub, the practice of splitting oak and chestnut lath to produce riven laths became popular early in the 15th century. Oak and chestnut make particularly good riven lath as they both contain natural oils, thus ensuring long life.

By the 19th century sawn lath started to be used, although there is no doubt that riven lath is stronger, and its textured surface and exposed grain affords a far better key.

Laths should be spaced about one centimetre (3/8″) apart, which is the distance between the top of your little finger nail and the underlying pad. (Spacing can be gauged simply by resting your little finger on top of the last lath to get a sufficiently accurate gap – you seldom see a true craftsman with a modern rule, measuring as he goes!) If the laths are fixed any closer, the first coat (or ‘scratch coat’) of plaster (often prepared from mature slaked lime putty and well graded, sharp aggregate ranging from up to 3mm through to fines) will not be able to pass through to form good nibs. Larger gaps will allow heavier nibs to form which are liable to break off, filling the void behind the laths.

In the 1999 edition of The Building Conservation Directory, the view was expressed that it was a mistake to think of a 1:3 lime/aggregate mortar mix as ‘standard’, as the proportions depend on the choice of aggregate and in particular its surface area and void. Suitable mixes can vary from beyond 1:3 down to 2:1, but if the proportions are not relevant to the aggregate the mix could well be totally unworkable. The same problem can also occur with haired mortar specifications as the weight to volume ratio of animal hair varies considerably from species to species. There is no point in specifying a particular weight of hair for a given quantity of lime mortar if the source and type of hair is not identified.

The importance of ensuring that sufficient hair is evenly distributed throughout the mix cannot be over-emphasised. In the past, apprentice plasterers often started their training by spending several weeks at a bench, beating bundles of hair with lengths of riven lath to break up any lumps and separate the fibres. Allowing lumps of hair into a plaster mix is almost worse than not putting any hair in at all, as the lumps have no binding power and create weak spots in the plaster, causing it to fail.

Where haired lime plaster is to be applied to a masonry background, the plaster relies largely upon suction for its bond, so it does not need as much hair as a plaster which is to be applied to lath, which relies almost entirely for its key on the nibs that protrude between the laths. Insufficient hair reinforcement in a plaster mix on lath will result in weak nibs and the risk of early failure.

Hair should be added to plaster just before spreading as the alkalinity of the lime attacks the protein in the hair. Millar, in his 19th century masterpiece Plastering Plain and Decorative (see Recommended Reading), writes of an experiment where haired lime plaster was stored in wet conditions for nine months: at the end of this period he found that the hair had ‘been consumed by the action of the lime’. Recent analysis of a failed ceiling plaster from an important civic building showed that, while the mortar mix proportions were adequate and the appropriate quantity of hair was almost certainly used, it had been introduced into the mix and ‘wet stored’ for several weeks prior to application. The weakened fibres were unable to maintain the integrity of the nibs which could then not support the weight of plaster, resulting in a failure that could have been catastrophic had the hall been occupied at the time.

For many thousands of years, lime plaster has been applied to masonry substrates and various forms of wooden lath. An increasing number of specifications now call for lime plaster to be applied to inappropriate backgrounds such as plywood, and a variety of alternative lathing materials have been introduced, including expanded metal and patent forms of lath product such as Riblath. Even chicken wire has been used. As there will be little or no bond with a parent background such as plywood, the integrity of the render is dependent on the choice of lath and the quality of fixing. All these attempts to bypass the proven traditional methods will fail unless the fixings are capable of taking the load and the mesh, or lath, is fixed in such a way that mortar can encompass the structure that is to support it. If expanded metal is specified, a system of spacers, such as simple timber batten, should be fixed behind the mesh to ensure that the plaster can surround the support.

When working on historic and, in particular, listed structures, repairs should ideally be carried out using similar materials to the original. Not only are they more appropriate to the historic character of the architecture, but they usually work better than modern alternatives, especially when used in conjunction with other traditional materials and construction techniques.