The Development of Portland Cement
A small scale lime slake
One of the earliest patents for a cement was taken out by Kendrick Edisbury in 1677 for a certaine sort of Plaister of an Extraordinary Hardness called ‘Glassis’. Little is known of the manufacturing process as this was not a requirement of patents at that time. However, from this it is clear that from at least the late 17th century, manufacturers in this country have been experimenting with a variety of hydraulic binders – that is to say, one which sets more rapidly by a reaction involving water – many of which were referred to over the years as ‘cements’.
There was considerable activity in the development of new alternatives to the simple
non-hydraulic lime mortars in the late 18th and early 19th century, partly fuelled by the desire to create the appearance of fine ashlar in areas of the country where brick was more readily available than stone. In 1796 an ‘hydraulic’ cement was patented by Mr Parker of London which he described as ‘Roman Cement’. In fact it had nothing to do with Roman origins except that it was of a similar dark brown colour to earlier Roman mortars which may have been non-hydraulic mortar converted to an hydraulic set by adding volcanic ash as a pozzolan. Parker was one of the first to experiment with ‘septaria’, nodules of which were found on the seabed off Harwich. This material was comprised of almost equal quantities of clay and calcium carbonate and the presence of traces of iron oxide gave the product a distinctive brown colour. Septaria was also found on the sea bed at Sheppey and was fired to produce ‘Sheppey cement’. A similar product was made from the same type of septaria found off the Isle of Wight and the remains of the Medina cement works can still be seen today. Another was Martin’s cement whichwas developed by Richard Green Martin of Lambeth in 1834 and Keen’s cement was patented in 1838 by J D Greenwood and R W Keen. Howes improved Keen’s cement followed a few years later. 1846 saw Parian cement patented by J Keating and products such as Robinson’s cement and Birmingham Waterproof cement were also developed at about this time.
Joseph Aspdin was also very busy during this period, developing ‘Portland’ cement which he patented in 1824. His product, which was supplied in barrels ready for use, must have appeared as an almost magical alternative to conventional lime mortars, as there was no longer a need to slake large lumps of burnt limestone in troughs of water or purpose built ponds, a process that had been carried out for the previous 10,000 years or more. It must have seemed as though the days of the ‘larry’ (a long handled ‘rake like’ device which was used to keep the lime moving during the slaking process) were at an end. This hot and gruelling task took a strong pair of arms and a great deal of effort to ensure that the lime, which would reach temperatures as high as 600ºC didn’t slake dry even though it was under water. Once slaked, it was usually left for several months to allow it to mature before being used. The whole process was very arduous and potentially very dangerous, as a man falling into a deep pit during slaking would probably be killed or at the very least, be blinded and suffer third degree burns.
In addition to being a manufactured material, Portland cement was simple to use: mix one part cement with three parts sand, add the water and the resulting mortar would
be hard by the next day. Compared to the considerable effort and time taken to produce a simple lime mortar, this alternative must have been most attractive. The rapid set also meant that, in wall construction, less time had to elapse before the next few courses of stone or brick could be laid, enabling construction to continue more quickly.
It is said that Aspdin’s newly patented Portland cement took its name from the observation that, when mixed with aggregate, it produced a mortar that was similar in colour to Portland stone. Much of the early use of Portland cement was as a binder in external renders and once rusticated or lined out to replicate stone, the comparison was noted. For the past 10,000 years or more, relatively pure limestone had been burnt at between 900ºC and 1,000°C as this is the natural limit when charging a kiln with wood. Coal only came into common use at the time of the industrial revolution when, as it has a higher calorific value than wood, higher temperatures were more easily achieved. Aspdin was probably aware that conventional lime mortars cured to varying strengths depending upon both the natural levels of clay in the limestone and that the process was also partially dependant upon temperature and humidity. A non-hydraulic lime mortar cures solely by absorbing carbon dioxide from the atmosphere as it dries. For the process to work, the temperature needs to be in excess of 10ºC and the humidity needs to be below 80 per cent, as otherwise the surrounding air will be unable to absorb the moisture.
Few limestone sources are pure calcium carbonate and most are contaminated to a greater or lesser degree with a variety of elements including forms of clay. It is the percentage inclusion of these elements that partly determines the hydraulicity of the burnt product. Most naturally occurring limestone vary from a few percentage of contaminate up to a maximum of about 20 to 25 per cent. In simple terms, the higher the percentage of clays and/or the higher the firing temperature, the greater the hydraulic reaction that could be produced. Aspdin decided to work with 100 parts relatively pure limestone and 45 parts clay. The materials were blended together, fired at 1,200ºC initially, and then crushed and fired at an even higher temperature before being ground to produce a fine powder.
Apparently his first experimental firings produced cement that took an initial cure so quickly that it would have been of little use on a building site. The addition of gypsum to the product acted as a retardant and cement was born. At twice the price of an equivalent quantity of lime, sales were slow and by 1850 the four factories were producing only 70,000 tons per annum. However, it was Aspdin’s son William who, in 1852, took out a patent for an improved method for manufacturing Portland cement and sales built up for the product which was much as we know it today. This was manufactured to a uniform standard, quite unlike the conventional process in which limestone was fired in small-scale kilns at slightly varying temperatures. Portland cement had the advantage over other cements in that it was strong, had good hydraulic properties and an ability to carry a large proportion of aggregate. While most lime mortar mixes revolved around 1:1 through to 1:3, cement mortars could be mixed with aggregate to a wider range of proportions. Early experiments carried out on cement showed that a mix of 1:1 was approximately 3.4 of the strength of neat cement and a mix of 1:5 was approximately 1.6 its strength.
While Portland cement has many advantages when designing modern structures, unfortunately it is not compatible with historic mortars, mainly because they were fired at the lower temperature of about 900ºC. The problem with firing limestone above this temperature is the generation of tricalcium aluminates and a combination of this and gypsum (an hydroscopic material) as a retardant in cement manufacture accounts largely for the minor movement that can occur in modern cement mortars, necessitating the inclusion of expansion joints within the masonry.
The strength and density of mortars made with Portland cement also cause problems where used on historic masonry. Strength is critical because to be suitable, the mortar must be weaker than the stone or brick it is used with, so that if stress in the wall leads to failure, fracture lines occur in the mortar, not in the brick or stone itself. Lime mortars are more flexible and can tolerate a greater degree of movement without failure. This in itself makes cement mortars and lime mortars disastrous bedfellows, particularly where a traditionally constructed wall has been repointed in a lime mortar. In this case the core remains flexible, but the surface is now rigidly bound by the new mortar, and seasonal movement causes stress in the surface, commonly resulting in spalling brickwork or stonework.
The density of the mortar or, to be more specific, its vapour permeability and porosity, are also critical because traditional construction methods rely on a porous surface to remain dry. In modern wall construction the cavity which separates the inner and outer leafs ensures that any moisture that enters the wall is unable to reach the interior. However in a traditional construction, whether a lime rendered panel of wattle and daub, fair faced brick or ashlar, or a solid masonry wall with a coating of render or limewash, the surface absorbs moisture readily, but it also dries readily, so damp does not penetrate far it into the wall. Preventing this natural evaporation with a dense waterproof coating of cement will trap moisture, risking damage by frost action, and salt crystallisation damage can occur wherever cracks in the coating occur.
Although most specialists are all too aware of the amount of damage which has been done by the inappropriate use of cement rich mortars since the First World War, the misuse of Portland cement was in fact already well established by the end of the Victorian period. Some of our most important medieval buildings were ‘restored’ and repaired using Portland cement, by architects of the Gothic revival like Street, presenting conservators today with some of their greatest challenges. Removing a tightly adhering mortar may cause more damage than it avoids.
The Aspdins’ discovery revolutionised construction. Their legacy, Portland cement, is an invaluable building material. However, like all materials, it is important to understand its properties and characteristics before using it, particularly where historic buildings are concerned.