About "Roman" Cement

This website is concerned with the history of manufacture of Portland Cement in Britain, but the earlier product, called Roman Cement forms part of the story, because many manufacturers who eventually made Portland cement began their business making Roman Cement.

Roman Cement was patented by James Parker in 1796 - see patent. An anonymous pamphlet of 1830 describes Parker's "discovery" of the product:

It was first discovered by the Rev. Dr Parker in the year 1796 and like many other of our most useful acquisitions, was purely accidental. When on a visit to the Isle of Sheppey, he was strolling along under its high cliffs on the northern side and was struck with the singular uniformity of character of the stones upon the beach and which were also observable sticking in the cliffs here and there. On the beach, however, the accumulation of ages, they lay very thick. He took home with him two or three in his pocket and without any precise object in view, threw one on to the parlour fire from which in the course of the day it rolled out thoroughly calcined. In the evening he was pleased to recognise his old friend upon the hearth, and the result of some unpremeditated experiments with it has been the introduction to this country of a strong, durable and valuable cement.

These "noddles of clay", as he called them, were septarian calcareous concretions, and are common in the London Clay which forms cliffs along the north coast of the Isle of Sheppey, on the south side of the Thames estuary. After the use of Parker's cement became popular, other sources of septaria or "cement stones" were sought, and concretions of a very similar character were found elsewhere.

Formation of Septaria

Warden Point Warden Point: Isle of Sheppey
A 45 m cliff of eroding London Clay. Reefs of septarian nodules have accumulated on the foreshore.

Nodule A nodule from Warden beach, sawn and ground flat. The main veins are macrocrystalline calcite, and are widest at the centre, narrowing towards the edge. In addition, there is a web of fine veins containing calcite or pyrite. The matrix is fine and fairly uniform, with most grains below 20 μm. The nodule was originally larger - the faces on the right are the result of pieces breaking away along veins. The veins contained sufficient traces of iron to weather to hematite red.

The coast of the Isle of Sheppey is the classic location. The London Clay is soft and wet, forming the base of gently undulating vales inland. Where a coast intersects it, a rapidly-eroding cliff results. The cliff erodes both by undermining by tidal action and by down-wash. In contrast to the soft clay, the septaria are quite hard, and when washed out of the surrounding clay, persist in the beach for a long time. Before large-scale harvesting of the beach nodules began in the nineteenth century, they lay in deep drifts along the more productive beaches. Since their use ceased towards the end of the nineteenth century, reefs of nodules have once again started to form along the Sheppey coast.

The formation of septarian nodules in clay appears to take place at an early stage in the deposition of the clay. The decay of organic material in the slowly consolidating mud releases carbon dioxide, which reacts with calcium in solution and calcium carbonate precipitates in the pore solution between incipient clay grains. In favourable circumstances this results in a fairly uniform fine-grained matrix with the required 60-70% calcium carbonate, with most grains in the 1-20 μm range. However, there is a tendency for the various minerals to grow into larger clumps here and there. The pore solution commonly contains significant amounts of magnesium, manganese and iron, and these replace some of the calcium in the carbonate phase.

The nodules become septarian at the point where the sediment becomes sufficiently deeply buried that the calcified nodule comes under vertical pressure and develops tensile cracks. These fill with pore solution and carbonates (and sometimes pyrite or apatite) crystallize in the void formed. Oriented growth of crystals in the gap then causes the crack to widen and propagate through the nodule, producing the characteristic pattern of cross-linked veins throughout the nodule.

The nodules of Sheppey are typically slightly flattened spheroids of 10-50 cm diameter. Their moisture content is typically only 2-3%, indicating the efficiency with which the pore-space of the clay is replaced with carbonate. This also explains the considerable hardness of the nodules - more characteristic of hard limestones.

The matrix part of a nodule similar to that shown here had analysis:

SiO215.89
Al2O37.78
Fe2O36.90
CaO34.55
MgO1.40
SO30.56
LoI95029.28
Na2O0.28
K2O0.42
SrO0.22
TiO20.60
P2O51.13
Mn2O31.00
Cl0.05

septaria map Coasts yielding septaria

Raw materials and plant locations

Parker set the standard for manufacture of the product, but only operated between about 1793 and 1797. He then sold the business to Samuel and Charles Wyatt, and the Wyatt family continued the operation until 1846. The patent lapsed in 1810, whereupon a great many others started making the product at many locations.

It would appear that Parker obtained his raw materials from the north and east beaches of the Isle of Sheppey on an ad hoc basis, but the Wyatts bought the Lordship of the Manor of Minster, giving them mineral rights on the foreshore, and entered into contracts with other Lords of Manors in Sheppey. It appears from litigation that they attempted getting nodules from above the high-tide mark - i.e. from the cliffs - but were prevented from continuing by the landowner. After 1810, new entrants secured their own mineral rights. James Frost, with Admiralty friends, had jumped the gun and started manufacturing at Harwich around 1807. Francis and White, with a works at Nine Elms (Battersea), had previously secured stocks and supplies of nodules, and commenced manufacture promptly on the patent expiry date. By 1811, works were in operation at Sheerness and Millwall, and at least a dozen works were in operation on the Thames by 1820. Also starting in 1811 was Atkinson's works at Sandsend, near Whitby. Numerous manufacturers commenced making cement all over the country in the period up to 1840. Not all these were legitimate Roman Cement manufacturers, but they used the name for local trade.

A large number of firms purporting to make Roman cement subsequently went on to make Portland cement, including:

Following large sales of Roman cement in northern France, it was first manufactured there at the plant of Dupont and Demarle at Boulogne in 1848, using septaria from the Kimmeridge Clay that outcrops in the cliffs to the north. They also went on to make Portland cement, from 1853.

Burning the nodules

Parker's manufacturing method consisted of burning the broken nodules at around 1000°C. The nodules were broken to about 40-80 mm. Although this was in part to aid rapid calcination, the main reason for breaking the nodules was to reduce the amount of neat calcite present. When broken, the nodules naturally break along the veins, which contain brittle calcite, allowing most of the contents of the veins to break away as dust, and leaving reasonably pure chunks of the uniform matrix. Thus the production of free lime in the product was reduced.

Parker used a bottle kiln rather than an open lime kiln. It is stated that the kiln was operated continuously, with equal volumes of coal and broken nodules added at the top and calcined material withdrawn at the bottom, with a residence time of 2-3 days. Coke was used as fuel when it became available in 1812. The energy consumption was probably in the range 12-15 MJ/kg. When cool, the calcined material was hand-crushed, then milled with flat stones, the product often being sieved prior to packing in casks. Parker situated his production plant at Northfleet because of the availability of a tidal mill and windmill for grinding. Packing was done quickly because of the extreme moisture-sensitivity of the powder, contact with damp air causing a considerable reduction in strength.

One ton of cement stone was said to yield about 21 bushels of cement. If the loss-on-ignition in the above analysis applies, then this implies a bulk density of 75.4 lb per bushel - a typical value for a lightly-burned cement. Hard burning was avoided because it led to longer setting time, due to progressive conversion of calcium aluminates into unreactive gehlenite (Ca2Al2SiO7). So although a high burning temperature was implied by the patent, the temperature actually used was typical of ordinary lime kilns, and vitrified material was thrown away. The Roman cement kilns used by G & T Earles at Wilmington, which continued in operation into the 20th century, were used to burn septaria of various origins (both Suffolk and Cleveland) and chemistries, and were small open kilns with no "bottle" extension.

During the century of its production, the technology of its manufacture did not evolve in any way. As with every "natural" cement, the manufacturer (and his customer) was entirely at the mercy of the vagaries of the raw material, and the final product quality was essentially a function of the accidental chemical make-up of the septaria.