Roman Cement Mortars in Europe’s Architectural Heritage of the 19th Century

Abstract

Natural cements calcined at low temperatures, so-called Roman cements, formed an important binder material in 19th century construction and facade decoration of many European cities and towns. As a rule, Roman cement mortars appear well preserved. In order to understand their range of composition and properties, a number of samples from historic buildings were collected and analyzed. Microscopic techniques including scanning electron microscopy were employed along with the assessment of physico-mechanical properties. The study shows that the most significant feature of Roman cement mortars is a wide range of differently calcined clinker relicts within each mortar. These unreactive particles, classified as being either overfired or underfired, obviously play an important role for the mortar properties. The binder relicts are composed of a number of phases in the system Ca-Si-Al-Fe. C2S and C2AS (gehlenite) are among the most frequent compounds, their grain size and microstructure depending on calcination temperatures. In particular, the clinkers produced at lower temperatures show nonequilibrium features such as solid solution systems and zoning by partial diffusion. The aggregates found in the mortars cover a wide range of mineralogical compositions reflecting local geological conditions. Cast and in situ applied mortars differ in the amount of aggregate, which is generally lower for cast elements. There is no correlation between the amount of inert material and the state of preservation indicated, e.g., by the occurrence of shrinkage cracks. Historic cast mortars show high compressive strengths at comparably low moduli of elasticity. The total porosity is frequently high. This contribution presents the above-mentioned properties and discusses them in terms of the excellent aging performance of the historic Roman cement mortars.

Natural cement and monumental restoration

“…The samples present an original aspect and show a very fine, beige coloured and homogenous texture. In most cases a sub-millimetre thick brownish oily-aspect layer underlines the surface, as previously observed by Weber [24] on roman cement mortars sampled from European monuments. Traditional repair mortars contain mineral additives (sand, stone powder, pozzolans and/or tile fragments) but the matrix of our samples does not contain any of these coarse inclusions.…”

Determination and imaging of binder remnants and aggregates in historic cement stone by Raman microscopy

“…Because of its rapid setting, the so‐called Roman cement, patented 1796 in England by James Parker, was in particular suited for the casting of building ornaments. A description of the many‐sided applications of Roman cement within a European context was given by Weber et al . Typical features of Roman cement mortars are different kinds of nonreacted binder phenograins (clinker remnants) embedded in the hydrated matrix.…”

“…Figure shows a typical C 2 S‐dominated binder remnant found in a thin‐section sample taken from the façade element mentioned earlier. Belite grains can be identified based on the typical ‘grape‐shaped’ arrangement of rounded colorless C 2 S crystals with yellowish to brown interstitial phases grown between them . In historical Roman cements, belite occurs in the two structural modifications α ′‐C 2 S and β ‐C 2 S, with an increasing proportion of the latter with higher (super‐optimal) calcination temperatures .…”

“…The second type of this class is fully hydrated but more dense and brittle than the surrounding cement matrix. (3) The third category of underfired phenograins exhibits the structure of the raw material (marlstone), contains widely unaffected quartz and occasionally fossil structures . These criteria allow a more detailed classification of the clinker remnants found in our samples.…”

“…We mainly attribute this to the first use of Raman microscopic imaging for studying Roman cement clinker remnants. As seen in the Raman map, only a few small γ ‐C 2 S crystals are present in a matrix mainly consisting of β ‐C 2 S. Therefore, local measurements with a spatial resolution on the order of 1 µm are needed to identify this polymorph, and the γ ‐C 2 S signals were presumably overwhelmed by the signals of the β ‐polymorph in the X‐ray diffraction (XRD) measurements applied so far …”