Elsevier

Construction and Building Materials

Volume 47, October 2013, Pages 1309-1316
Construction and Building Materials

Statistical analysis on ancient mortars: A case study of the Balivi Tower in Aosta (Italy)

https://doi.org/10.1016/j.conbuildmat.2013.06.026Get rights and content

Highlights

  • The statistical analyses distinguished mortars of different rebuilding periods.

  • The role of variables was studied to select the most suitable technique.

  • Different degradation agents in different heights and facades were found.

Abstract

This study proposes a first approach to the characterization of the composition and degradation products of historical mortars and their classification through chemometric analysis. This work also aimed at comparing the usefulness and applicability of analytical techniques that are commonly used for mortar examination. Calcimetry, Thermogravimetric Analysis (TGA) and Differential Thermal Analysis (DTA), Ionic Chromatography (IC) and Fourier Transform Infrared Spectroscopy (FT-IR) were used to analyse a number of samples collected from the quadrangular Tower of the Balivi Complex in Aosta (Italy). Cluster Analysis and Principal Components Analysis (PCA) were carried out on the analytical results, and the accordance and the relevance of the results were evaluated. The role of the variables on the samples description and their correlation were evaluated. The results showed the division of the samples into groups that were probably due to the presence of different mortar types manufactured in different periods. Furthermore, an irregular degradation pattern on the four facades and at different heights of the tower was highlighted.

Introduction

Mathematical and statistical methods are widely used in studies on cultural heritage. Chemometric analyses support dating and provenancing of the raw materials of archeological and historical objects such as glasses [1], [2], [3], [4], [5], [6], ceramics [7], [8], [9], [10], organic binders [11] and stones [12]. A chemometric approach allows the user to extract relevant information from a large number of data set that has to be managed after a research study [13].

Despite the widespread application of chemometrics to the conservation of cultural heritage, only few a statistical studies aimed at mortar characterization have been made. All of these studies were only concerned about the grouping and the classification of samples, and nothing has been published that directly compares the various analytical techniques that can be used for characterizing mortars.

Rampazzi et al. [13] used PCA to investigate specimens and samples of historical mortars. A PLS (Partial Least Squares regression for one y-variable) was needed to create a statistical method for estimating the binder/aggregate ratio (B/A). Using a SIMCA method (Soft Independent Models of Class Analogy), Musumarra et al. [14] distinguished Gothic and Flemish mortars used for mural paintings in Chiaravalle Abbey. The test results from thermal analysis, mercury porosimetry and mechanical strength analysis were used by Moropoulou et al. [15], [16] for the classification of historical mortars in Byzantine and Ottoman monuments using PCA. Principal component analysis (PCA) was also used by Giuffrida et al. [17] to distinguish between painted plasters and wall covering plasters. Alvarez et al. [18] proposed a method based on hot hydrochloric acid attack of mortars to carry out a separation of the binder and the aggregate, while Sanjurjo-Sánchez et al. [19] used a multivariate statistical analysis of the chemical data obtained by X-ray diffraction and neutron activation analysis for the differentiation of mortars groups from the Santa Eulalia de Bóveda temple (Spain).

This study is thus one of the firsts chemometrical studies on mortars and is, in particular, the first chemometric study comparing analytical techniques that are currently used for their examination.

This research aims to classify and date mortar samples collected from the quadrangular tower of the Balivi Complex in Aosta. The information acquired was compared with archive data of previous studies [20], [21], [22] to distinguish the different mortar mixtures and to determine the various rebuilding periods and restorations.

This study also aims to evaluate the validity and the actual need for the most common analytical techniques used for mortar characterization such as Calcimetry, Thermogravimetric Analysis (TGA), Differential Thermal Analysis (DTA), Ionic Chromatography (IC) and Fourier Transform Infrared Spectroscopy (FT-IR) [23], [24], [25], that in some cases provide comparable information.

A statistical investigation of the results of such analytical techniques should result in a reduction in the number of analyses needed to characterize mortars.

Section snippets

Mortar samples

Sixteen samples in total were taken using a scalpel from all four facades at different heights of the Balivi tower (Fig. 1). The sampling locations are shown in Fig. 2. The samples were finely ground using an agate mortar, to obtain a particles size of about 230 mesh. The powdered samples were analyzed by Calcimetry, TG/DTA, IC and FTIR.

The calcimetry was performed using the UNI 11140:2004 procedure [26].

For TG/DTA analysis, a Pyris Diamond TG/DTA Differential Thermal Analyzer was used, with

Variables

All the variables for the study of mortars composition were used for both Q-mode Cluster Analysis and PCA using a matrix of 16 objects by 8 variables. In this study case only the first two principal components were considered, as their cumulative explained variance reached 79% (52% and 27% for the first and second principal component respectively). This value is considered suitable for an adequate description of the whole data set and, therefore, for correct interpretation of the results.

The

Conclusions

This study offers a new approach to the statistical study of ancient mortars. Based on their composition, Cluster Analysis and the PCA showed a diversification between the mortar samples. This division is probably due to different manufacturing periods for the mortars, coinciding with documented rebuilding works that affected the Balivi Tower Complex in Aosta after the Barbarian Invasions.

For the first time this investigation evaluates the role and agreement between variables acquired after

References (26)

Cited by (10)

  • Earthen mortars from Cremona (Northern Italy): The evolution throughout centuries of a manufacturing tradition

    2016, Construction and Building Materials
    Citation Excerpt :

    EDS analyses of CF were also treated with a statistical approach based on the principal component analysis (PCA). This approach is extremely helpful to unravel complex geochemical data [8,34,35]. Two data matrixes of some compositional parameters (SiO2, Al2O3, K2O, FeO, MgO, CaO, Na2O, SO3, P2O5 and Cl) were prepared.

  • Characterization of XVIIIth century earthen mortars from Cremona (Northern Italy): Insights on a manufacturing tradition

    2015, Materials Characterization
    Citation Excerpt :

    Chemical compositions were collected considering a 100% oxide content on a H2O- and CO2-free basis. In order to highlight any major variation in the set of compositional data, a statistic approach (cf. [25,26]) was applied using Principal Component Analysis (PCA). Untreated fragments of mortars were platinum coated and high magnification images (InBeam mode) were collected with a working distance of 5 mm to show the morphologies of aggregates at the nanometer scale.

  • Statistical analysis of the physical properties and durability of water-repellent mortars made with limestone cement, natural hydraulic lime and pozzolana-lime

    2015, Construction and Building Materials
    Citation Excerpt :

    In fact, if the solution is able to enter inside the porous structure, then the mechanical strength and the internal cohesion of the specimens determine the resistance. Our results agree with what obtained by other authors [12,18] on the exposure of water-repellent mortars to soluble salt solutions, thus indicating that the treatment of physical data regarding weathered mortars might be a promising application of PCA, as demonstrated also by [30]. The increasingly frequent approach of statistical modelization of the composition and different properties by linear regression models was fruitfully adopted in this study.

View all citing articles on Scopus
View full text