The Importance of Complex Aggregates in Historic Renders and Repair Mixes by Peter Ellis

Posted 6th March 2020

There are three critical factors to be considered when designing a mortar:

  • The selection of the appropriate binder
  • The selection of the correct aggregate(s)
  • The volumetric proportions of binder(s) to aggregate(s)

The choice of aggregate is to a large extent governed by the type of binder selected. This
fact is often overlooked or ignored, and this is without doubt contributing to some
examples of poor performance of mortars, plasters and renders.

The History

Prior to 1800 the vast majority of mortars for all applications were based on a lime binder. These traditional limes, whether used as quicklime for hot mixes, or as lime putty varied considerably due largely to the type of limestone burnt and the firing conditions and temperature in the kiln which ranged from 8500C to about 9000C.

One factor that they had in common was that they were impure. The nature of the impurities is crucial. Alumina and silica which contain phases which can react with lime are technically most important. The precursors of reactive alumina and silica are clay minerals which are present in many limestones. In addition to reactive alumina and silica other impurities included:

  • Unburned Coal or Wood
  • Coal ash or charcoal
  • Unburnt ( unconverted ) limestone
  • Partially converted limestone
  • Reconverted calcium carbonate ( limestone converted to calcium oxide or quicklime
    in the kiln which has reacted with the carbon dioxide in the kiln and re-carbonated )
  • Quartz silica
  • Unreactive clay minerals
  • Flint ( in the case of chalk limes)

It was not uncommon to find the majority of these impurities in one lime type or source. Traditional lime binders therefore already contained complex aggregate.

Early Aggregates

These limes were then mixed with almost anything available locally to produce a mortar,
plaster or render. These materials included:

  • Earth or clay
  • Quartz
  • Felspars
  • Flint / Chert
  • Coal / Coal Ash
  • Crushed rock such as granite, dolerite, basalt and sandstone
  • Crushed limestone and chalk
  • Alabaster
  • Mixed aggregates such as Beach or River sand and glacial deposits
  • Crushed brick and tile

These aggregates were generally added as dug. They were rarely washed although river sands for example were naturally low in fines or clays content. They were however often crushed or sieved down to a particle size appropriate for the intended use. Many of these materials had a positive or beneficial physical and chemical effect on the performance and durability of traditional mortars, and this is the critical issue when
considering renders, or external plasters, as this application demands a great deal of a mortar.

Many of the potentially disadvantageous properties of these as dug aggregates are not relevant in traditional limes for these reasons:

  • Traditional limes do not contain potentially damaging hydraulic compounds capable of reacting with, for example, sulphate leading to sulphate attack
  • Traditional limes are flexible and therefore capable of absorbing any slight expansionary reactions or forces which may occur
  • Traditional limes are not strong and do not rely on compressive strength for durability
  • The two groups of material which have a positive chemical effect are:
  • Calcareous or limestone aggregates
  • Pozzolanic materials.

Calcareous aggregates

By calcareous aggregate, I mean calcium carbonate present either as impurities in the lime binder or as limestone in the aggregate.

The reaction series which is critical to the performance of historic mortars is carbonation, and not the hydration of calcium silicates and aluminates.

To understand this, it is necessary to understand the chemistry of carbonation, and this is;

I believe more complicated than generally thought.

Carbonation is understood in the oversimplified reaction of:

Ca(OH)2 + CO2 → CaCO3 ( Lime + Carbon dioxide → Calcium carbonate)

As moisture is necessary, the reaction should be written:

x.H2O + CO2 + Ca(OH)2 → CaCO3 + y.H2O

In Fact

The first reaction is that of CO2 + H2O to form H2CO3 (carbonic acid). Reactivity increases in the presence of acidity in rainwater. It is this carbonic acid which reacts with lime (on the surface) of mortar to produce calcium carbonate.

Ca(OH)2 + H2CO3 → CaCO3 + 2H2O

The depth of this surface reaction depends on many factors including time, permeability
of mortar, and relative humidity

Later:

Carbonated lime on the surface of mortar reacts with carbonic acid to form calcium
bicarbonate.

CaCO3 + H2CO3 → Ca(HCO3)2(aq).

Both the calcium bicarbonate and the carbonic acid permeate into the mortar where they both react with lime to produce calcium carbonate:

Ca(OH)2 + H2CO3 → CaCO3 + 2H2O

Ca(OH)2 + Ca(HCO3)2 → 2CaCO3 + 2H2O

In a Traditional Lime, or mortar containing limestone aggregate:

As long as sufficient CO2 is available to stabilise the bicarbonate in solution, which is likely to occur in the long term, calcium carbonate present as limestone aggregate, or as unconverted, partially converted, or reconverted binder material, will in turn dissolve in the carbonic acid to produce calcium bicarbonate. This bicarbonate will in turn react with lime to form calcium carbonate.

CaCO3 + H2CO3 → Ca(HCO3)2

Limestone Aggregate + Carbonic Acid → Calcium bicarbonate

Ca(HCO3)2 + Ca(OH)2 → 2CaCO3 + 2H2O

Calcium bicarbonate + Lime → Calcium carbonate + water

Mortars containing limestone carbonate within in addition to on the surface.

The particular forms of limestone (carbonate), and the silicate impurities present in
traditional limes may facilitate this reaction.

Pozzolanic Materials

A pozzolana is defined as a material, which is capable of reacting with lime in the presence of water at ordinary temperatures to produce cementitious compounds. They occur naturally as earths and clays of volcanic origin. Artificial pozzolanas are made by burning certain clays and shales at suitable temperatures and include crushed brick and tile as well as the metakaolins. PFA ( pulverised fuel ash ) is pozzolanic as are certain
diatomaceous earths and some coal ashes.

Both the Greeks and Romans understood that certain volcanic deposits when added to lime and sand produced a mortar which possessed superior strength and set under water, and was therefore hydraulic. They used to transport these pozzolanas large distances, but if none was available the Romans made use of crushed bricks, tiles and pottery which had a similar effect. There are many examples of extremely durable mortars and lime
concrete from the Roman period both in Italy and England.

This technology was lost in Europe after Roman times and not re-established until the seventeenth century in England. There are no volcanic ash deposits suitable for use as pozzolanas in Great Britain, but there are many examples of 17th and 18thC renders where brick particles and powder were deliberately included. I have recently analysed a remarkably durable yet relatively soft 18thC render sample from Frogmore House in the
Windsor Great Park which had significant quantities of both brick and coal ash deliberately added.

To my knowledge there are no pozzolanic aggregates in Great Britain although some would disagree with this and are of the opinion that, for example, the metamorphic ring around Dartmoor yields reactive aggregate. A major cement manufacturer has measured expansion in Portland cement: Thames grit mortars and has, I believe, concluded this to indicate a pozzolanic reaction, although I have not seen this data.

The lime-pozzolana reaction is a progressive reaction between reactive silica and alumina and lime to produce calcium silicate hydrates and calcium aluminate hydrates similar, but not the same, as those produced in hydraulic limes and Portland cement.

Particle shape and grading

It is often considered that sharpness is a desirable characteristic in an aggregate as it ensures a better bond between binder and aggregate. Sharp aggregates produce less workable mixes and therefore require more water which tends to reduce strength and increase shrinkage.

In my analysis of historic mortars, I have found no correlation between sharp aggregates and durability. Similarly, aggregates in traditional mortars were rarely well or evenly graded, and if they were it was accidental. Again I have seen no correlation between particle grading and
durability. The shape of aggregate particle is however, important in top coat roughcast renders. Historically round pebbles or peas were included in the thrown top coat, and these as well as being visually pleasing, helped throw rainwater off the wall. Both limestone and sandstone rounds are still available.

To sum up so far, if I were asked to design a lime render for a traditional stone or brick building that would stand the best chance of proving durable for, say 300 years, the ingredients would be:

  • An impure lime putty produced in a coal-fired kiln
  • Coal ash sieved to 5mm down
  • A pozzolanic material such as metakaolin or reliably pozzolanic brick particles and powder
  • Limestone sand (not limestone dust) sieved to 5mm down
  • Quartz, flint or sandstone sand sieved to 5mm and well graded
  • Limestone rounds if the top coat is roughcast

The render must then be protected with an exterior grade limewash with added tallow, linseed oil, or casein.

So far, I have been talking about aggregates appropriate for use with traditional lime putty and non-hydraulic limes.
One has to be much more careful and cautious when selecting aggregates for use with hydraulic limes, as the range of aggregate type is more limited.
Hydraulic limes vary considerably ranging from the weakly hydraulic such as the English blue lias HL2, and French St Astier NHL-2 to the eminently hydraulic varieties, some of which hardly contain lime and I consider these grades to be best described as (weak)
cements.

Clearly, these binders do have advantages in wet and aggressive environments and for winter working, and they are appropriate binders for repairing or matching many 19th and early 20th C renders. The manufacturers of these products generally recommend that only washed sands and
grits are used as aggregate. These sands, grits and gravels are derived from the weathering of rocks and are in general composed of the more resistant materials which have been able to withstand for a long period the destructive effects of weather and movement by glacier, river, or the sea.

Hydrated hydraulic limes, in particular the more hydraulic grades, are very different to non-hydraulic lime. In the production, an impure limestone containing certain clay minerals is heated to above 9000C and a series of reactions occur in the kiln of which three are crucial:

  • Reaction of CaO with SiO2 to form Calcium silicates.
  • Reaction of CaO with Al2O3 to form Calcium aluminates
  • Dehydroxylation of clays.

These materials subsequently react during hydration to give a product which is hydraulic i.e. the solids are strongly flocculent (they coagulate) and gain strength even under water.

The cementitious compounds present in hydraulic limes make them vulnerable to deleterious reactions with some aggregate types, and this significantly restricts the range of aggregates to select from.

These reactions include:

  • Sulphate attack
  • Coal and clinker expansion
  • Alkali-aggregate (alkali-silica) reaction

In addition, the durability of these materials depends largely on the strength of the mortars. An unwashed sand where each sand grain is coated with clay interferes with the bond between binder and sand particle and reduces strength. Significant quantities of any fine material, for example pigment, will reduce strength and the addition of large quantities of fine limestone dusts is of particular concern. This is a common practice with lime putty mortars to achieve an appropriate colour but this will be disadvantageous in hydraulic mortars. This is particularly relevant when
limestone has been mechanically ground to produce the powder as the grinding process may well produce reactive silica from the fine quartz impurities.

Sulphate Attack

Hydraulic lime renders are vulnerable to the soluble sulphates of various bases. Sodium sulphate for example present in groundwater and some unwashed aggregates will react with any free lime to produce calcium sulphate. This more than doubles the solid volume and expansion and cracking occurs.

Calcium sulphate will react with hydrated calcium aluminates, particularly tri-calcium aluminate, present in many hydraulic limes, to form calcium alumino-sulphate. This again more than doubles the solid volume and expansionary damage results. A hydraulic lime render failure involving massive expansion in the south of England has been attributed to sulphate attack. This largely explains why only washed sands and additives with a low sulphate content should be used with hydraulic limes (and cements).

Coal and clinker expansion

Coal, coal ash and clinker should never be added to hydraulic lime mortars. These materials, depending on the source and grade can contain sulphates and sulphides. There is another potentially damaging reaction which consists of a progressive expansion of the mortar. This depends on the nature of the coal burnt and is associated with the swelling of unsound coals due to their absorption of oxygen, and this expansion of mortar will be greater if the mortar is wet. British coals and clinker are particularly prone to this, and I am aware of a hydraulic lime render failure in the UK due to this reaction.

Alkali-aggregate reaction (aka alkali-silica reaction)

This is a deleterious expansive reaction between alkalis ( sodium and potassium hydroxides ) in the binder ( and possibly the aggregate) and reactive silica in the aggregate. These alkalis are present in cement and to a lesser extent in hydraulic limes, and the problem is that of two incompatible materials. Fortunately, this reaction is rare in the U.K. but there are authenticated cases in Denmark, Norway and Sweden, where the sedimentary sands and gravels contain porous opaline silica and flint. Great care is needed when using aggregates of geological types which might contain such constituents and of which no service experience is available.

This reaction has occurred in the U.S.A. with opal and cryptocrystalline volcanic rocks  and metamorphic schists. Cryptocrystalline quartz is present in some sands in the Plymouth area of Devon, and metamorphic schists are found in Cornwall, Scotland and parts of Wales. The use of any of these aggregates with hydraulic lime binders is (in my opinion) best avoided.

I have experience of one hydraulic lime render failure in England in which I believe an alkali-aggregate reaction was the cause. 

Discussion and Questions

Alick Leslie said that historic documents referred to volcanic ash for use as a pozzolan being sourced in or near Edinburgh but that the source could not now be located. He also stated that there had been much use of metamorphic schists as aggregate in Scotland without trouble. Tim Palmer  explained that there may be a difference between American and Scottish schists. Paul Livesey agreed that he was not aware of problems with Scottish schists, but said that they had found the Greywacke ( ancient sandstones) of the Borders region to contain reactive silica in the form of very fine reactive quartz. He also confirmed that the Thames grits and gravels contained reactive amorphous silica (flint but apparently not opaline) which is causing expansive damage in cement-based mortars. Jersey grits which do contain opaline flint were also causing problems. (Tim Palmer advised there was no such thing as opaline flint; – this is a term commonly used in industry but apparently not correctly!)

Marianne Suhr asked if unfired clays were pozzolanic. The answer to this is almost certainly no, but it does depend on clay type. Certain diatomaceous clays are mildly pozzolanic in their natural state. The HR Supra brick used by English Heritage in the Smeaton trials is a fired diatomaceous clay from Denmark. It is not surprising therefore that the fired material is reactive and reliably pozzolanic.

Jeff Orton asked about marble dust as aggregate, a common practice in Italy in lime plasters. There is no doubt that this has a beneficial effect. As marble is metamorphic limestone, it has been thought that this may be a pozzolanic reaction. In my (limited) analysis of this however, I have not found evidence for this. Tim Palmer said that the physical nature of marble would promote the formation of calcite. It would be of great interest if this could be shown to be the case as this would explain one of the puzzles of the lime business.

Have a question?
Contact one of our specialists