Building for Eternity

the History and Technology of Roman Concrete Engineering in the Sea

C.J. Brandon, R.L. Hohlfelder, M.D. Jackson, J.P. Oleson, R.L. Hohlfelder, M.D. Jackson

• Publisher: Oxbow Books
• Year: 2014
• ISBN:
  9781782974215
• Topic: Scienze sociali
• Subtopic: Archeologia

One marker of the majesty of ancient Rome is its surviving architectural legacy, the stunning remains of which are scattered throughout the circum-Mediterranean landscape. Surprisingly, one truly remarkable aspect of this heritage remains relatively unknown. There exists beneath the waters of the Mediterranean the physical remnants of a vast maritime infrastructure that sustained and connected the western world's first global empire and economy. The key to this incredible accomplishment and to the survival of structures in the hostile environment of the sea for two thousand years was maritime concrete, a building material invented and then employed by Roman builders on a grand scale to construct harbor installations anywhere they were needed, rather than only in locations with advantageous geography or topography.This book explains how the Romans built so successfully in the sea with their new invention. The story is a stimulating mix of archaeological, geological, historical and chemical research, with relevance to both ancient and modern technology. It also breaks new ground in bridging the gap between science and the humanities by integrating analytical materials science, history, and archaeology, along with underwater exploration. The book will be of interest to anyone interested in Roman architecture and engineering, and it will hold special interest for geologists and mineralogists studying the material characteristics of pyroclastic volcanic rocks and their alteration in seawater brines. The demonstrable durability and longevity of Roman maritime concrete structures may be of special interest to engineers working on cementing materials appropriate for the long-term storage of hazardous substances such as radioactive waste.A pioneering methodology was used to bore into maritime structures both on land and in the sea to collect concrete cores for testing in the research laboratories of the CTG Italcementi Group, a leading cement producer in Italy, the University of Berkeley, and elsewhere. The resulting mechanical, chemical and physical analysis of 36 concrete samples taken from 11 sites in Italy and the eastern Mediterranean have helped fill many gaps in our knowledge of how the Romans built in the sea. To gain even more knowledge of the ancient maritime technology, the directors of the Roman Maritime Concrete Study (ROMACONS) engaged in an ambitious and unique experimental archaeological project – the construction underwater of a reproduction of a Roman concrete pier or pila. The same raw materials and tools available to the ancient builders were employed to produce a reproduction concrete structure that appears to be remarkably similar to the ancient one studied during ROMACON's fieldwork between 2002-2009.This volume reveals a remarkable and unique archaeological project that highlights the synergy that now exists between the humanities and science in our continuing efforts to understand the past. It will quickly become a standard research tool for all interested in Roman building both in the sea and on land, and in the history and chemistry of marine concrete. The authors also hope that the data and observations it presents will stimulate further research by scholars and students into related topics, since we have so much more to learn in the years ahead.

Chapter 1

The Technology of Roman Maritime Concrete

J. P. Oleson and M. D. Jackson

1.1. Introduction

The central purpose of this book is to present literary, archaeological, and analytical data concerning Roman concrete structures built in the sea, with particular focus on the highly specialized marine concrete developed for that purpose. We have brought together and translated all the Greek and Latin literary sources that describe maritime concrete and its applications, the materials, formwork, and tools used to produce it, and, to an extent, the ancient interpretations of the geological origins of those materials. Careful interpretation of these texts in combination with results of archaeological, experimental and analytical investigations provides important information on Roman practical knowledge and engineering procedures for building in the sea. We have also put together catalogues of known Roman concrete structures constructed in the sea (Chapter 6) and of the remains of the formwork used to create these structures (Chapter 8). Although these catalogues undoubtedly are incomplete, the first provides a general idea of the geographical spread of the technology of maritime concrete construction, while the second documents both widespread uniformity and local innovation in the design of Roman concrete formwork. While the materials used in Roman concrete bridge footings and lakeshore structures are undoubtedly relevant to the topics discussed in this book, with the exception of the bridge at Chalon-sur-Saône (pp. 219–20) and a concrete embankment at Lake Nemi (p. 127), we have chosen to focus on the marine structures that we had the opportunity to sample.

The bulk of the book, however, is a report of the activities and results of the Roman Maritime Concrete Survey (ROMACONS), directed by Brandon, Hohlfelder, and Oleson between 2002 and 2009 (Chapter 4), and of the scientific analysis of the resulting concrete cores carried out by Jackson, Vola, Gotti, Bottalico, Cucitore, and Stern, and researchers at the University of Naples (Chapter 7, Appendices 3–4). Over the seven years of fieldwork, the ROMACONS team took 36 cores (totalling 36.55 m in length) from 11 Roman harbour sites and one fishpond in Italy, Greece, Turkey, Egypt, and Israel. The cores are described in Appendix 3. A wide variety of physical, chemical, and microstructural analyses was carried out on the cores, producing the results presented in Chapter 7. A synthesis and historical appreciation of the results of the research is presented in Chapter 8.

This book is not intended to be a general introduction to Roman concrete engineering, or a history of how continued innovation in the mixing and placing of concrete affected the evolution of Roman building design on land. Numerous surveys of these topics already exist (e.g. Blake 1947; Lugli 1957; MacDonald 1982; Lamprecht 1996; DeLaine 1997; Taylor 2003; Lancaster 2005).

Given the multidisciplinary character of this book, which involves ancient literature, archaeology, and the physical sciences, terminology can become problematic. A glossary of technical terms that frequently appear has been provided in Appendix 1, in the interest of avoiding repetitive explanations, or laborious periphrasis. Volcanic ash is the product of an explosive pyroclastic eruption; it is composed of glass and crystals derived from magma, or molten rock, and particles of rock, mainly lavas broken from the underground edifice of the volcano. Tuff is the rock that forms when volcanic ash lithifies and consolidates through the development of natural mineral cements. A pozzolan is a siliceous and/or aluminous material, named after ash from Pozzuoli (ancient Puteoli), which reacts with lime or lime-based compounds in the presence of moisture at ordinary temperatures to produce compounds with cementitious properties (Massazza 1988). Above all, it should be noted that the term pozzolana is used only rarely in this book, given its widely ambiguous meanings in both Italian and English. The term “pozzolanic additive” is also avoided in this book, since this has specific applications to modern cement technologies that do not pertain to the ancient concretes. In the modern literature concerning Roman concrete, for example, the term pozzolana can be used to indicate a type of powdery, pumiceous, incoherent volcanic ash erupted from the Campi Flegrei volcanic district that surrounds the Bay of Pozzuoli at the northwest sector of the Gulf of Naples (Fig. 1.1, 7.2), pumiceous volcanic ash from elsewhere along the coast of the Gulf of Naples, or scoriaceous volcanic ash found in and around the city of Rome. Instead, we use the more straightforward “pumiceous ash pozzolan” or “mortar containing pumiceous volcanic ash.”

Fig. 1.1. Puteolanus pulvis (pozzolana) from a quarry near Baia.

Pozzolanic materials add durability and long-term strength to modern cementitious materials, even in maritime environments. In antiquity, the most common pozzolans were pyroclastic rocks – mainly poorly consolidated volcanic ash or glassy tuffs. Vegetable ash pozzolans were also sometimes used on a large scale (Lancaster 2012: 146). The altered volcanic tuffs, or trass, of the Rhine region were finely ground and used as pozzolan in the mortars of Roman concrete structures at Cologne during the first and second centuries (Lamprecht 1996: 61, 75, 87; Elsen 2006). Ground-up potsherds and brick also produce pozzolanic reactions with lime, and these were frequently used by the Romans, producing opus signinum for floors and cistern linings (Italian, cocciopesto; Blake 1947: 322–23; Lancaster 2005a: 58–59).

In Latin, Vitruvius’ term for the pumiceous, poorly-consolidated volcanic ash that crops out “in the vicinity of Baiae and the territory of the municipalities around Mount Vesuvius” in the northwest sector of the Gulf of Naples was pulvis, “powder” or “dust” (De arch 2.6.1). This term refers to its finer grain size distribution, as compared with the granular scoriaceous ash or excavated sands (harenae fossiciae) of the region around Rome. Vitruvius thus indicates that the powdery ash used in first century BC came from either the Flegrean Fields near Puteoli or the Somma-Vesuvius volcanic districts (Fig. 7.2). The term, Puteolanus pulvis, or “dust (or powder) from Puteoli,” occurs in two of the three passages by ancient authors that mention pulvis (Seneca, Q Nat 3.20.3; p. 26, Passage 14; Pliny, HN 16.202; pp. 26–27, Passage 15). In Pliny HN 35.167 (p. 27, Passage 16) the phrase is a pulvere Puteolano. Vitruvius does not attach a locative adjective, but simply states pulvis. The mention by Vitruvius and other Roman authors of Puteoli and the coastline of the Gulf of Naples as sources of pulvis for marine concrete has led many modern scholars to assume that all the pumiceous volcanic ash used in Roman marine concrete was sourced from this region. While the new literary, archaeological, and geological investigations described here have led the authors of this book to regard this as a reasonable hypothesis, our analytical results and those of previous studies are seldom perfectly conclusive. As a result, the association of the pyroclastic materials in the ancient concretes – mainly pumice and tuff – with the Gulf of Naples volcanic deposits is often expressed in a tentative manner. This approach may surprise readers accustomed to the confident attributions seen in many archaeological publications (see below pp. 2–5), but the reasons for this caution are explained in Chapter 7.

We refer to the material that is the focus of this book as “Roman maritime concrete” or “Roman marine concrete,” rather than “Roman hydraulic concrete.” The latter is a general, generic term that refers to concretes that harden by reacting with water and form a water-resistant product. Romans did not use kiln-fired cements as we know them. Instead they relied on the reaction of hydrated lime with volcanic ash to produce stable binding cementitious hydrates. Most ancient Roman concretes used on land, as well as that in the sea, remain intact when saturated in water, and even develop new cementitious phases.

1.2. The unique character of Roman maritime concrete

The earliest synthetic lime mortars, simple mixes of slaked lime and quartz sand, appear in the archaeological record in the Near East as early as 12,000 BC, and these were applied to architectural uses by 10,000 BC (Gourdin and Kingery 1975; Kingery et al. 1988). Probably not by accident, and possibly in connection with early ceramic production or metallurgy, it was discovered that heating limestone to 800–900° C produced a caustic alkaline powder, calcium oxide (CaO). The principal component of most limestone is calcite, or crystalline calcium carbonate (CaCO₃). During calcination in kilns, calcium carbonate releases CO₂ gas and decomposes to calcium oxide (CaO), called lime (or quicklime). When quicklime is mixed with water, or “slaked,” an exothermic hydration reaction takes place that produces hydrated lime (Ca(OH)₂), or portlandite. Vitruvius described this reaction as it occurred during the slaking of lime with fresh water to form putty for the volcanic ash mortars of architectural concrete structures (De arch 2.6.3; pp. 17–19, Passage 7).

When slaked lime putty is mixed with quartz sand, the portlandite carbonates in the presence of atmospheric CO₂ to form a calcite cement binder. The resulting mortar develops some compressive strength and resistance to shrinking and cracking. This type of mortar is non-hydraulic, and it may deteriorate during long term saturation in water after having set. Nevertheless, simple lime mortars provided adequate strength and water resistance to serve as plaster on floors, walls, and roofs, and for the lining of cisterns throughout the Mediterranean area for many centuries. Plasters were widely used during the Bronze Age (Shaw 2009). By the Hellenistic period, similar mortars were also used in the Aegean world to bind rubble walls and to provide a smoothjoint between dimensioned stone blocks (Theophrastus, On Stones 65; Martin 1965: 422–33; Hellman 2002: 94). By the late Republican era, Romans had developed careful techniques regarding both the design of their limekilns and the selection of limestone for calcination (Cato, Agr. 38; Vitruvius, De arch. 2.5.1–3, pp. 16–17, Passage 6; Adam 1994: 65–76; DeLaine 1997: 88–89, 111–14).

Several authors have stated that hydraulic mortars in certain Classical and Hellenistic Greek structures at Santorini (Thera), Athens, and Rhodes were formulated deliberately with volcanic pozzolans (Martin 1965: 424, 432: Koui and Ftikos 1998; Stamatopoulos and Kotzias 1991). The pozzolan is usually said to be the siliceous volcanic ash of the Santorini eruption of approximately 1600 BC. The chronology of some of the structures involved, however, is poorly established, and the components of the mortar, as far as can be determined, have not been subjected to thorough analysis. It is clear that the local pumiceous ash on Santorini was used by local island builders in mortars and plasters from the Archaic through the early modern period. The ash may have improved cementitious properties, and it was applied to both structures meant to contain water and those that were not. This suggested to early archaeologists that local and Roman builders alike did not understand that Santorini ash could produce a hydraulic mortar, that the ash may have been added as inexpensive bulk filler, and that it was thus unlikely to have been exported (Wilski 1909). This perspective underestimates the empirical expertise of both the local builders and the Imperial age Roman builders, who began to develop their great technological expertise with high performance concretes in Rome during the late first century BC (Jackson et al. 2009, 2010, 2011; Jackson and Kosso 2013). Siddall (2000: 340) believes there is no evidence for the export of pumiceous volcanic pozzolans from Santorini or Melos during the pre-Roman period. Although the ROMACONS project has not detected pumiceous volcanic ash from Santorini or Melos in the Roman harbour structures that were cored, further research is needed on the proposed use of Santorini ash at Athens and Rhodes, and on the possibility that there was a modest export trade to other Aegean sites during the Imperial period. The new results from mineralogical studies and trace element signatures of pumices in the Roman maritime concretes presented here provide new insights into Roman builders’ selections of volcanic ash pozzolans for the maritime concretes (Chapter 7).

From the fourth century BC onwards at some sites in the Aegean world, crushed brick was added to mortars used to line cisterns (Martin 1965: 432). The large-scale production of pozzolanic mortars for applications in water-saturated environments, however, began at some point in the third or second century BC, most likely in the landscape of the Campi Flegrei volcanic district (Latin: Campi Phlegraei = Phlegraean Fields), whose central crater forms the Bay of Pozzuoli in the northern sector of the Gulf of Naples. Most of the ancient literary sources that mention pozzolanic mortar concern this region, and the highly valued pumiceous ash pozzolan, Puteolanus pulvis, was and is still excavated in the volcanic craters near ancient Puteoli and Baiae (see Strabo and Pliny, below; also Maffei 1949; Lugli 1957: vol. 1, 394–401; Lancaster 2005a: 54–58). Blake (1947: 346) dates the beginning of harbour construction at Puteoli to 199 BC, but the remains of the arcuated pier visible until the early twentieth century probably belong to the Augustan period (Döring 2003: 47). The history of the earliest concrete, therefore, remains fraught with uncertainty. Although Blake (1947: 328–30) mentions literary evidence for various construction projects in Rome in the third or second century BC that might have used concrete, she states that the Temple of Concord erected in Rome in 121 BC “furnishes the earliest concrete of which the date is sure.” Controversy now surrounds the identification, function, and age of the so-called Porticus Aemilia, on the left bank of the Tiber River near the Aventine Hill (Lancaster 2005a: 5). Lugli (1957: vol. 1, 409; cf. also pp. 375–85) uses the traditional date of 193 or 174 BC, but new analyses by Tucci (2012) suggest a later date for the opus incertum construction, perhaps in mid-second century BC. Geochemical and petrographic investigations of the mortars of late Republican concrete structures indicate that builders experimented with various ash deposits of the Roman landscape through the late first century BC, until they standardized a specific scoriaceous ash formulation during the Augustan era (Van Deman 1912a; Jackson et al. 2010, 2011). Many aspects of the earliest concrete structures must have been experimental, resulting in early failure, or repairs, replacement, and incorporation in later structures, and further analytical studies of Late Republican concrete structures, such as those described in Jackson and Kosso (2013) are needed. Vitruvius states (De arch. 7, preface 18) that he wrote the De architectura to fill the gap left by earlier Roman architects who had not written down the principles of their work. It seems, however, that the most important advances in concrete construction in architectural settings occurred in Rome in the mid- to late first-century BC, based on both observations of structures (Van Deman 1912a) and analytical investigations of concrete materials (Jackson et al. 2010, 2011), and this may be true of maritime concrete construction as well (Jackson et al. 2012).

In Rome, volcanic pozzolans were excavated first from alluvial deposits in the city, and then from the mid-Pleistocene Pozzolane Rosse pyroclastic flow erupted from the nearby Alban Hills volcano (Jackson et al. 2010). The granular scoriaceous ash has a grain size distribution with a large proportion of sand-sized ash particles, described by Vitruvius as harenafossicia, or “excavated sands” (pp. 15–16). At Portus, for example, the pozzolanic mortars of the marine structures in the harbour proper appear to have been made with pulvis imported from the Gulf of Naples (see Figs 7.10, 13), while associated structures on land were made with local dark gray and reddened scoriaceous harena fossicia (Delaine 2001: 248). Vitruvius described the characteristics of both materials and distinguished their different functions in structures on land and in the sea (De arch. 2.6.6; Passage 7; pp. 17–18). These descriptions undoubtedly were based on the practical experience of first-century BC builders (Jackson and Kosso 2013). In fact, pulvis and harenae fossiciae in the ancient mortars have quite different chemical and mineralogical compo...