Engineering Rome

Understanding Roman Concrete

Nigel.Lyons Nigel.Lyons Sep 16, 2013


The purpose of this article is to inform readers about topics necessary for understanding Ancient Roman concrete. Concrete preforms vital roles in nearly all aspects of public works including infrastructure systems and buildings. Versatility, strength, and workability make concrete a universal construction material that has become a foundation to everyone’s daily lives. The Ancient Romans first discovered this technology over 2000 years ago. Ancient Roman knowledge and skill with concrete can be observed first hand through massive concrete structures that stand today such as the Pantheon (Rome, Italy 126 AD, Figure 1). In today’s environment, massive engineering feats like the Three Gorges Dam (Hubei province, China 2012, Figure 2) or the Burj Khalifa Tower (Dubai, UAE 2009, Figure 3) exemplify the substantial abilities that modern concrete has. Concrete is now the building blocks for modern civilization.

Pantheon F1.jpg Three Gorges F2.jpg Burj Tower F3.jpg
Figure 1: Pantheon (Google images). Figure 2: Three Gorges Dam (Google images). Figure 3: Burj Khalifa Tower (Google images).

Most of the knowledge of Ancient Roman engineering comes from a person known as Vitruvius. He lived somewhere around the time of 80 BC to 15 BC. He wrote a substantially detailed work known as the Ten Books on Architecture that survived the multiple sacks that Ancient Rome experienced. The majority of today’s knowledge on Ancient Roman engineering comes from Vitruvius’s work.

Concrete was never a single scientific discovery. It developed slowly through a long process of trial, luck, and keen observations. Concrete technology actually advanced overtime on two separate occasions. The Ancient Romans had developed consistent concrete technology around the start of the Roman Empire in 42 AD. However, starting in the 3rd century AD, the fall and decline of the Roman Empire forced the knowledge of concrete to be forgotten until the late 18th century. After its rediscovery, the redevelopment process continued again into what we consider modern-day concrete.

Background and Definitions
Before further discussion on ancient and modern concrete, a few clarifications must be made on the definitions and [[#|meanings]] of the various words associated with concrete. The following section will provide insight including cement, hydration, and pozzolan, elements that will prevail throughout the article.

Cement vs. Concrete

Modern concrete is essentially artificial rock that is created through the mixing of cement (which acts as the [[#|bonding agent]]), water, and aggregate. Modern cement is generally known as Portland cement, the primarily consumed cementing material worldwide. Aggregate refers to anything the builders put inside their [[#|concrete mix]] besides cementing materials. This filling material usually includes other course and rough rocks such as sand and gravel of varying sizes. The bonding and strength developments during curing, when the concrete hardens, due to a process called hydration. During hydration, chemical changes take place and the aggregate particles are essentially glued together to form an artificial rock. Although hydration sounds like the addition of water, the process actually occurs while the concrete cures and ironically the concrete does indeed “dry-up”, in the sense that it transitions from a wet rocky-paste into an actually solid structure (PCA, 2013).

Hydraulic Cement vs. Non-hydraulic Cement (aka Mortar)

Hydraulic cement is defined as any cement that undergoes hydration (reaction mentioned above) to cure and harden into a desired structure. All Portland cements are hydraulic cements. Hydraulic cements are sometimes referred to as being “water resistant” because they can cure in wet or submerged environments and do not deteriorate with contact with water. This property attributes itself to the cement’s reaction with water that actually initiates the curing into a solid structure; in fact, prolonged contact with water allows the concrete to continue to increase in strength. On the other hand, non-hydraulic cement, referred to as mortar, is a lime based paste that hardens through a reaction with the CO2 in the atmosphere. Thus, when exposed to water as a paste, mortar will be unable to cure and harden into a solid mass.

Modern vs Ancient Pozzolan

Modern pozzolans, or pozzolanic materials, include any supplementary cementing materials that contribute to hardening due to the hydration of Portland cement. Pozzolans can include the following (PCA, 2013):

  • Natural pozzolanic ash: naturally produced volcanic ash buried relatively close to the surface.
  • Fly ash: ash produced from the combustion of coal.
  • Blast-furnace slag: nonmetallic silicate (rock like material) separated from metals during smelting or refining.
  • Silica fume: ultrafine silica powder that is a by-product of elemental silicon and ferrosilicon production.

In Ancient Rome, Vitruvius eludes that they used two materials with pozzolanic properties. The first one, Vitruvius referred to as pitsand. Pitsand was the natural sand-like pozzolanic ash that was found by digging large open holes close to the surface. Pitsand is also known as pozzolana because it was found in great abundance in areas surrounding the city of Pouzzuoli (or Puteoli). The second pozzolan material the Ancient Romans used was crushed brick, which was burned clay. Although today we consider both burned clay and natural pozzolanic ash as pozzolonic materials, in Vitruvius’s writing, pozzolana only refers to the volcanic ash known as pitsand. In the following sections, pozzolana will refer to the volcanic ash pitsand.

History of Cement

The following section will give a general overview of major events and steps in ancient and modern time that led to the construction material we consider today as cement. Cement’s history involves ancient development, a technological regression and forgotten knowledge, and its modern rediscovery.

Ancient Development of Cement

One of the first ancient building materials was clay. Clay’s surface and shallow lithosphere abundance combined with its workability and cohesive properties made clay a simple and primitive building material. Prehistoric clay could be used in three ways: walls could be constructed with the raw, earthy material by compacting piles either by hand or with wood planks. Secondly, clay could also be mixed with rocks and compressed with temporary wood planks; walls of this nature from 200 BC still exist today in Spain (Moore, 1995). The previously mentioned use of clay parallels methods used in both ancient and modern concrete placing methods. Namely, compressing mixed materials and using wood boards to retain forms. Lastly, sundried clay bricks could be stacked and layered; Moore tells that the first evidence of clay bricks can be seen in excavations as far back as 8000 BC in the Middle East and 4000 BC in Iraq. Walls, dwellings, and other structures could be constructed by shaping and sun-drying clay and then stacking them on top of each other. Eventually clay was burnt in large kilns as opposed to sun-drying. Kilns are oven-like structures meant to burn clay bricks and lime that were eventually used by the Ancient Romans to create stronger bricks.

The introduction of lime mortar was the next major step in the development of ancient concrete. Limestone is a sedimentary rock composed of mostly small grains of settled marine skeletal structures such as coral or tiny shelled organisms. Limestone could be burned in large kilns to produce quicklime. Quicklime could then be mixed with water, a process known as slaking, to produce a pasty-like mortar material through a chemical reaction (discussed in detail a later section). This lime mortar can be seen being used in walls as far back as 2000 BC in central India. Archeologists also uncovered early use of lime mortar in the construction of Minoan foundations in Prehistoric Greece, Crete 1700 BC. It was discovered that many foundations were typically composed of previous structures and buildings that had collapsed from earthquakes. Lime mortar would be incorporated into the remaining rubble to solidify and develop the foundation for the next building that would be constructed on top of the rubble of the previous (Moore, 1995).

Lime mortar was used by Romans as both a construction material as well as a plaster meant for aesthetic relief due to its brilliant white finish, appearing much like marble. Vitruvius tells that this plaster was known as stucco and was typically a combination of water, lime, and various types of sands. Powdered marble would be added for a more brilliant white finish. The exact date that Romans came to their understanding of lime remains disputed and largely unknown; however, lime based mortar was prevalent in the end of the 3rd century BC (Adam, 1994).

The next significant advancement in Roman lime-mortar was the addition of crushed tiles and brick, composed of burned clay. With the burned clay-lime mix, Romans had discovered their first hydraulic cement. Romans used this water resistant mortar, known by the Ancient Romans as opus signinum, to line water infrastructure such as the many aqueducts, used to carry water over large distances into the city, and the many castella, large cisterns used to hold, filter, and distribute water throughout the city (Vitruvius, c. 15 BC).

The most significant change in Ancient Roman lime mortar was the accidental addition of pozzolana, the sand-like volcanic ash known as pitsand to the Ancient Romans. Lime mortar was often mixed with sand. In fact, Vitruvius accounts of three types of sand, each with varying properties. These included river sand, marine sand, and pitsand, sand that was excavated from the ground in pits surrounding Naples. These sands frequently were interchanged in the lime mortar and eventually, the Ancient Romans realized the pitsand, actually strengthened the mortar and allowed for underwater curing. With this, Romans had discovered their second form of hydraulic cement. The pozzolana-lime based cement was used to construct large maritime and structural works that needed the upmost highest bonding quality.

The first account of this pozzolana-lime cement was in Pompeii dating around 3rd century BC. Archeologist found a wall constructed of poor quality cement and rubble mixture as seen in Figure 4. Rubble refers to a mixture of rocks and stones of varying size and shape that makes up the aggregate, or filler, in Ancient Roman concrete. In 199 BC, Puteoli harbor works was an early example of good quality hydraulic cement. It is considered good quality because the structures remain today after surviving the harsh conditions of the sea. The geographic location both Pompeii and Puteoli had abundant sources of pitsand deposits which explain why the first pozzolana-lime hydraulic cements arrived in those areas. It took about a century before the Ancient Romans made this connection and started using the pozzolana-lime cement in the city of Rome. By the Augustan period and the start of the Roman Empire in 42 BC, guided processes and construction practices had been refined and standards were well established (Figure 5). This was confirmed through observations of the detailed similarities between buildings arising in that period. For example, the temple of Saturn and the temple of Divus Julius had standardized concrete foundation and wall construction (Moore, 1995). By this time, Ancient Roman hydraulic cement, both pozzolana-lime based and crushed brick-lime based, was fully developed.

F4 Pompeii Wall.jpg F5 Rome Contruction Ex.jpg
Figure 4: Typical early pozzolan-lime-rubble concrete wall from Pompeii 3rd century BC (Pompeii). Figure 5: Looking into a typically advanced broken Ancient Roman wall (c. 42 BC (Ostia Antica).

Technological Regression and Forgotten Knowledge

Following the Hadrina period (138 AD), most of the great Roman infrastructure had been completed. Over the course of the next century, damages sustained from earthquakes, flooding of the Tiber River, and fires slowly accumulated and were left mostly unrepaired due to economic failure and poor leadership. By the 3rd century AD, the decline and fall of the Roman Empire initiated and ran its course over the next few centuries. The exact causes are heavily debated and it can be assumed that a combination of many complex factors were involved: economic failure involving a conquest based economy and a large unsustainable army; environmental degradation of natural resources and over harvesting; a declining population due to reducing water supply and spreading disease; barbarian military advancement and growing invasion pressures; poor leadership and political crisis. Edward Gibbon, an English historian (1737-1794), even theorized that the fall of the original Roman paganism and the growth of Christianity transformed social outlook on worldly living due to the promises of heaven. Regardless of the causes, the declining empire extinguished engineering advancement and opened the door to a period of technological regression on hydraulic cement and concrete.

Three direct causes led to losing the knowledge of hydraulic cement. Firstly, the poor economic state and lack of funding halted major construction projects. With little construction occurring for over a century, the demand of knowledgeable craftsmen and contractors vastly decreased. Secondly, the barbarian sack of 410 AD caused the few remaining craftsmen and contractors to flee to the countryside. Once out of the city, these families continued on substance living in which knowledge of hydraulic concrete quickly became unwarranted. Lastly, as the Middle Ages progressed, political and economic focus moved away from Rome and into Northern European cities such as London, Paris, and Cologne. Thus, the natural pozzolanic ash that was vital for the hydraulic cement was geographically absent (Moore, 1995). As a result, the knowledge of hydraulic cement was thus forgotten for over a millennium.

Modern Rediscovery and Advancement of Cement

Although natural pozzolana and volcanic ash deposits lay buried in southern Italy, limestone was very abundant throughout the world. As a result, lime-based mortar mixed with rubble and brick remained a primary construction material throughout the Middle Ages. It was not until 1756, when an English civil engineer named John Smeaton was tasked with the rebuilding of the Eddystone lighthouse that hydraulic lime cement was rediscovered. Smeaton experimented heavily on lime with many different admixtures and eventually discovered a hydraulic lime by combining clay with quicklime (the product of burned lime discussed ina later section). The clay Smeaton used contained multiple impurities that shared similar chemical compounds with the pozzolanic ash the Romans used. With this, Smeaton had developed the first hydraulic cement in over a millennium.

Smeaton’s clay-lime cement opened to the door to the advancement of modern cement. In 1796, James Parker, a cement manufacturer, showed that by grinding the burned lime into powder, the gel making process was greatly accelerated and improved. The finely powdered form of quicklime and clay has a large surface area to volume ration and thus increases the total surface area in which hydration can more readily take place with the additions of water. Since then, it has been common practice to produce cement in a finely powdered form that can be mixed with water and hydrated on site.

The effects of pozzolanic ash when combined with lime mortar were rediscovered by a French laboratory researcher who focused on construction. C.J. Vicat experimented on a variety of chemical processes with lime. One of the materials Vicat used was volcanic ash from regions surrounding southern Italy. In doing so, Vicat was able to observe the chemical reactions that allowed lime mortar to harden and strengthen underwater, thus rediscovering the pozzolanic effects of volcanic ash in hydraulic cement.

The final step in the development of what we consider modern cement was the establishment of Portland cement. In 1824, Joseph Aspdin combined clay and limestone particles through a heating process known as sintering. Sintering involves heating the materials to a special point. This allows the diffusion of calcium oxide particles into the silica molecule structure, effectively fusing the particles together. Sintering essentially melts and fuses the edges of the particles together without melting the core of the particles (Moore, 1995). This material, known as clinker, is later grounded into a powder and packaged for transportation and storage or use. At this point, Aspdin coined Portland cement in name of the grey-colored stone found on the Isle of Portland, England.

Reaching the 20th century, high temperature heat treatment of the various cement components and fundamental chemistry on clay impurities and their effects was understood by the many cement manufacturers. At this point, manufactures were at the beginning of using various supplementary materials to modify the characteristics and performance of specific types of cement. With this, Portland cement was a well-established construction material and modern concrete was implemented in multiple engineering projects in most of the developed world.

Modern Concrete

Modern cement is produced through an industrialized process. The main raw materials are clay and limestone that has been mined and grounded into very fine powder. The powdered raw materials are mixed together in specific proportions and heated to 2600F in a kiln to produce small chunks of a material called clinker. The clinker is then grinded into powder and a mineral used to control setting called gypsum is added. The mixture is then sealed in moisture proof bags to be stored and transported to the construction site. Once on site, the cement is mixed with water and aggregate and then placed into the various forms demanded by the structure. Hydration occurs with the addition of water that bonds the cement gel and aggregate together. With this, concrete is formed as the mass strengthens and hardens into an artificial, rock-like structure.

Reactions in the Kiln

Summarized below in Table 1 are common compounds involved with cement chemistry. Note that the clay and limestone rows neglect other minerals in the materials and only correspond to the raw materials that directly contribute cement production. Additionally, the reactions that take place in cement production and hydration require long chemical formulas. Thus, cement chemists take on a new abbreviation scheme to simplify reactions known as Cement Chemistry Notation (CCN). Many compounds below will be reference throughout the remainder of this article.

Table 1: Cement Compounds Names and Abbreviations

Common Name Chemical Name Chemical Formula Cement Chemistry Notation (CCN)
Limestone* Calcium carbonate CaCO3 n.a.
Clay* Silicon dioxide SiO2 n.a.
Clay* Aluminum(I) oxide Al2O n.a.
Quicklime (lime) Calcium oxide CaO n.a.
Alite Tricalcium silicate 3CaO*SiO2 C3S
Belite Dicalcium silicate 2CaO*SiO2 C2S
Aluminate phase Tricalcium aluminate 3CaO*Al2O3 C3A
Ferrite phase Tetracalcium Aluminoferrite 4CaO*Al2O3*Fe2O3 C4AF
Gypsum Calcium sulfate dehydrate CaSO4*2H2O n.a.
Calcium silicate hydrate Calcium silicate hydrate 3CaO*2SiO2*4H2O C-S-H
Hydrated lime Calcium hydroxide CaO*H2O CH

Decomposition of Raw Materials

At temperatures up to about 2300F, the raw materials are broken down and chemically react with lime to create intermediate compounds used to form the final clinker product.

  • At 1000F, water is removed.
  • At 1750F, calcination occurs and limestone loses its CO2. The gas is driven into the air, leaving the lime in a highly reactive state.
  • As lime is produced, belite is produced. Aluminate and ferrite phases also start to form.
  • At intermediate temperatures, sulfates combine with calcium to form a sulfate liquid phase. (Winter, 2013)

Alite Formation
As temperatures reach above 2300F, aluminate and ferrite phases liquefy. These liquid phases (including the sulfate phase), contribute to ion mobility and promotes fusing, at high temperatures. However, these intermediate phases separate and are not present in the final clinker material (Winter, 2013).

  • In addition to liquid phases, some liquid belite is included as well as dissolved lime and silica compounds.
  • At 2550F, belite is transformed into alite. Additional alite is formed from the free lime and silica.
  • Alkali sulfate, and some akali chloride, liquid phases evaporate and are passed back up the kiln process and again condense to liquid in cooler sections. The alkali phases continue to recycle and aid in the formation of other compounds.

Clinker Cooling and Processing
After the formation of alite, the clinker is taken away from the burner and cooled. During cooling, the main liquid phases (aluminate and ferrite) crystalize. The gypsum is then added to the clinker and the mixture is grinded into the desired particle size. At this point, the cement is sealed in moisture proof bags and ready for hydration.


Three main reactions take place during hydration of the cement (Winter, 2013).

  1. Sulfates and gypsum minerals immediately dissolve to produce alkaline, sulfate-rich solutions. This occurs minutes after water is added.
  2. Aluminate (C3A) forms aluminate-rich gel when contacted with water. This gel in turn reacts with the sulfate solution produced from the previous reaction to form small rod-like crystals called ettringite. This reaction causes the paste to go dormant, a time in which the paste is most workable. At this time, the cement-aggregate paste is placed into the position it will cure. The dormant stage generally lasts a few hours but workability decreases as the paste stiffens.
  3. Alite (C3S) and belite (C2S) start reacting with the water to form calcium silicate hydrate (C-S-H) and calcium hydroxide (CH).

Moore (1995) explains the hardening and strength gain during hydration to the diffusion, surrounding and expansion into pores on the molecular level, of the C-S-H gel into the components of the remaining cementing materials, namely CH, and the aggregate adjacent to the curing cement. The C-S-H gel then hardens into a lattice of very small interlocking fibers and plates. This is the primary process that results with the strength gain and curing into a concrete structure.

Strength that builds up from the C-S-H occurs during the third stage and can take place over long periods of time, depending on the rate that all the C3S and C2S is hydrated. Generally, 90% strength gain is usually observed within 28 days of hydration; however, in the right environments, concrete can continue to gain strength over time. Gotti et al. (2008) determined that a replicated, Vitruvian-formulated, hydraulic concrete with a six month cure time had a compressive strength of about 670 psi. They also tested original Ancient Roman hydraulic concrete that was dated to the 1st century as having a compressive strength to 1160 psi. The Ancient Roman sample has had over 2000 years to hydrate.

Ancient Roman Concrete

Ancient Romans created hydraulic cement with the combination of lime, pozzolan, and water. Vitruvius conveys the mix proportions used for varying structural functions. For example, harbor works and bridge piers would need the highest quality volcanic ash pozzolan found in Naples and a mix of two parts pozzolan and one part lime was required. Mortar, used for either brick or concrete cement, could be made from either three parts volcanic ash with one part lime or with a combination of one part lime, one part crushed brick, and two parts standard sand. Construction functions such as flooring surfaces and aqueduct/cisterns that required less structural demand are mixed with pozzolan-lime rations of 5:2, thus having less lime, and in turn less calcium, in the mix. Less calcium would result with a smaller quantity of C-S-H gel that attributes strength to the concrete. The Ancient Romans realized this and conserved the relatively expensive lime and uses high proportions only when necessary.

Vitruvius explains that the Ancient Romans would layer and compact the hydraulic cement with rubble between brick work to construct their structures. Figures 6, 7 and 8 all exemplify the Ancient Roman concrete work as well as the various common brick patterns they used: opus incertum, opus reticulatum, and opus testaceum respectively. Note that the occasional strip of flat rectangular brick that is uniformly spaced seen in Figures 9. These were used to occasionally ensure level and uniform brick placing.

F6 Inc.jpg F7 Ret.jpg
Figure 6: Opus incertum is composed for varyous rock chunks mixed with cement. Notice that there is no real brick facing. Opus incertum was one of the earlier concrete construction methods (Ostia Antica). Figure 7: Opus reticulatum was taped rectangular prisms that were placed on the outside of a wall with cement. Concrete would then my placed inside the brick work (Ostia Antica).

F8 Test.jpg F9 Flat Lines.jpg
Figure 8: Opus testaceum was a brick pattern that arrived later. Again, the outer brick work would be laid, then the concrete would fill in the open space (Ostia Antica). Figure 9: Combination of bricks for quality control during construction and potentially aesthetic reasons (Ostia Antica).

Additionally, the brickwork could later be coated with a white plaster for aesthetic relief, known as stucco, or to be a canvas for a fresco, a type of mural painting. The stucco was composed of three layers of varying amounts of fine sand with pasty slaked lime. The final layer to be viewed could also be mixed with powdered marble for increased brilliance. Figure 10 and 11 display the many layers of stucco applied to walls as well as a section of a fresco.

F10 Stucco Work.jpg F11 Fresco layers.jpg
Figure 10: Stucco outer covering with concrete rubble on the inside (Pompeii). Figure 11: Stucco layered over the brick work that conceals concrete rubble. The outer most layer of stucco differs slightly with the addition of powdered marble and the application of a Fresco (underground Ancient Roman home).


The Ancient Romans got their lime source (also known as quicklime, CaO) from raw limestone through a process called calcination. The limestone is heated to 1700F where water and CO2 gas is released leaving CaO. At this point, the Ancient Roman cement process deviates from that of modern methods. The CaO, in a chalky rock form, is mixed on site with water, a process known as slaking. During slaking, moderate amounts of heat is released as the lime is hydrated to produce calcium hydroxide, or slaked lime (Ca*2OH and CH in Cement Chemistry Notation). The slaked lime paste at this time is a non-hydraulic mortar that will only cure when exposed to the atmosphere.

Non-hydraulic lime curing and hardening occurs through a process known as recarbonation, essentially the opposite reaction as its formation. As the atmosphere evaporates the water, Ca*2OH becomes CaO again. The final curing reaction occurs as CO2 reenters the paste and combines with CaO to produce CaCO3. The creation and use of non-hydraulic mortar can be summarized as making raw limestone into a workable paste that will turn back into limestone material through a mirrored reaction after being shaped to its designated function.

The Romans used the lime mortar primarily for what Vitruvius referred to as stucco work. He explained that mixing the slaked lime with sand and sometimes powdered marble would produce a durable and brilliantly white finish when applied in layers on any wall, vault, or ceiling. Large murals, known as frescos would later be applied for aesthetic and artist expression.


The Ancient Romans later discovered that hydraulic cement could be produced with the addition of a finely powdered pozzolan material during the slaking process. Ancient Romans had two pozzolan materials: the volcanic ash, found in deposits surrounding Naples, and the burned clay products that had been crushed into powder. Figures 12 and 13 show the general geological surrounds of both Rome and Naples. From Figure 12, we can see that areas surrounding Rome were abundant with raw building materials used for aggregate (limestone, sand, and tuff) and also the clays used to be burned and made into the second pozzolan material. The areas surround Naples, as seen in Figure 13, demonstrate the abundance of volcanic ash.

Geography of Rome.jpg Geography of Naples.jpg
Figure 12: Geography of Rome. Figure 13: Geography of Naples. Notice the large circular element just east of Naples. The object is Mt Vesuvius, the major source of volcanic ash on the lowlands.

Moore (1995) has also summarized the chemical make-up (percentages) of the pozzolan materials used by the Ancient Romans in Table 2. Notice the similarities between the ancient pozzolan materials: silica (SiO2), alumina (Al2O3), ferric oxide (Fe2O3) and lime (CaO).

Table 2: Ancient Roman Pozzolan Materials

Pozzolan Material SiO2 Al2O3 Fe2O3 CaO Ignition Loss Other
Volcanic ash 55.8 19.2 4.0 3.6 4.6 12.8
Burnt clay 59.2 18.1 8.5 3.3 1.5 9.4

The volcanic ash and burnt clay provides the amorphous silica (a non-crystalline, more reactive state of silica), alumina, and ferrite materials required for the cementing processes. When combined with hydrated or slaked lime (CH), conditions allow the materials to hydrate, thus creating a strong bond with the aggregate.

Observations and Thoughts the Comparison Between Ancient and Modern Concrete


Modern concrete is composed of 11% Portland Cement and 67% aggregate and 22% air and water (PCA, 2013). Modern aggregate consists of gravel, crushed stone, and sand of various sizes. A good variety of aggregate sizes allows the concrete to better bond into and stronger, interlocking structure as it cures. Gotti et al. (2008) determined through Vitruvius that Ancient Roman concrete was a 65% cement-mortar paste and 35% aggregate with an ambiguous and liberal application of water with the cement paste. The aggregate was generally made of porous rocks including tuff and sandstone was well as finer aggregate such as sand and occasionally broken/disposed pottery pieces. Figures 14, 15, 16, and 17 all illustrate the various aggregate found in Ancient Roman concrete.

F14.jpg F15.jpg
Figure 14: Showing examples of porous rocks (dark) and tuff (lighter) commonly used in aggregate (Pompeii). Figure 15: Harder igneous rock was also used as aggregate (Pompeii).

F16.jpg F17 Tile Agg.jpg
Figure 16: Small pieces of various volcanic and sedimentary rock are mixed along with the cement before being applied with the rubble stones (Ostia Antica). Figure 17: Occasion tile pieces are visible today in the ancient concrete (Ostica Antica).

A curious note is how the cement-aggregate ratios contrast so highly between ancient and modern concrete. Ancient Roman concrete had a much higher cement paste composition of 65% while, almost reversed, modern concrete actually has 67% aggregate. One possibility explaining this discrepancy can be seen in the figures above. The Ancient Romans lacked a good aggregate-size gradient, or variation. Ancient Roman aggregate sizes were mostly all medium sized chunks of rock and stone with relatively small sand grains and rock flecks. To compensate for the large spaces between the aggregate, the Ancient Romans had to apply more cement paste to generate a good structural bond.


Table # below summarizes the ancient pozzolan materials in addition to the general chemical composition of modern Portland cement. At first glance, significant differences exist between all chemical components. Most notably is the CaO, content (Moore, 1995; Winter, 2013).

Table 3: Ancient Pozzolan vs. Portland Cement

SiO2 Al2O3 Fe2O3 CaO
Ancient pozzolan 57.5 18.7 6.2 3.5
Portland cement 21.5 5.2 2.8 66.6

This large difference can be accounted for my recalling how Ancient Romans used their cement. Recall that the Ancient Romans first made a lime based mortar through slaking. The hydrated lime was then mixed with the pozzolan materials on-site directly before placing. The pozzolan material in Table 3 is before the addition of the lime material (CaO), thus explaining the only trace amounts of CaO.

On the other hand Portland cement, through the sintering in the kiln, combines the lime material (CaO) into the powder like cement that would later be hydrated with water on-site before placing. If the lime component (CaO) is neglected from the Portland cement and the chemical composition recalculated, the cements appear much more similar. This is illustrated below by Table 4. Now the chemical compositions of the cementing materials are much more alike. This chemical composition is what allows the complex process of hydration to occur in both ancient and modern cement.

Table 4: Ancient Pozzolan vs. Portland Cement Adjusted for Lime

SiO2 Al2O3 Fe2O3 CaO
Ancient pozzolan 57.5 18.7 6.2 3.5
Portland cement (neglecting lime) 47.6 19.5 16.5

The only remaining difference is the increased amount of the ferrite in modern cements. Hydration will not be vastly affected with the change in ferrite because hydration, and strength, is more dependent on the C-S-H compound forming from calcium and silica components. If the ferrite is negligible during hydration, it is probably a factor in the sintering of modern cement, the other major difference in modern and ancient cement processes. It is postulated that the liquid ferrite phase in the kiln is increased to further promote ion mobility and fusing speed during the sintering process. Thus, the abundant ferrite compound is present for increasing the efficiency of industrialized cement production in the modern world.

Final Thoughts

Concrete is an artificial rock that has become an essential building material in the modern world. The technology was first developed by the Ancient Romans over 2000 years ago partially through geological luck and partially through strong observations. The volcanic ash settling in the southern regions of Italy provided the Ancient Romans with pozzolan, an essential ingredient for the hydration and curing of concrete. The technology was then forgotten and lost with the fall of the Roman Empire starting in the 3rd century AD. Only in the past 250 years has modern concrete redeveloped. Although modern and ancient concrete technology, especially cement processing and mixing methods, differs in many ways, the fundamental chemical composition remains mostly the same. The only differences in ancient and modern concrete appear in placement methods, aggregate, and small cement mixture variation due to modern industrialization. One can only wonder what modern civilization would be like today if Ancient Roman engineering survived and continued to advance.


Adam, J. (1994). Roman Building: Methods and Techniques. (A. Mathews, Trans.). Abington, England: B.T. Batsford Ltd. (Original work published 1989)

Gotti, E., Oleson, J.P., Bottalico, L., Brandon, C., Cucitore, R., & Hohlfelder, R.L. (2008). A Comparison of the Chemical and Engineering Characteristics of Ancient Roman Hydraulic Concrete with a Modern Reproduction of a Vitruvian Hydraulic Concrete. Archaeometry, 50,576-590. doi: 10.1111/j.1475-4754.2007.00371.x

Moore, D. (1995). The Romand Pantheon:The Triumph of Roman Concrete. Mangilao, Guam: David Moore.

PCA, Portland Cement Association. (2013). Concrete Basics. Retrieved from

PCA, Portland Cement Association. (2013). Supplementary Cementing Materials. Retrieved from

Vitruvious. (1914). The Ten Books of Architecture. (M.H. Morgan and A.A. Howard, Trans.). London, England: Humphrey Milford Oxford University Press. (Original work published c. 15 BC)

Winter, N., (2013). Cement hydration. Retrieved from

Winter, N., (2013). Portland cement clinker – overview. Retrieved from

Follow us

Don't be shy, get in touch. We love meeting interesting people and making new friends.